diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_2pimofs7erzvi.abi3.so b/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_2pimofs7erzvi.abi3.so
deleted file mode 100755
index fef48439dbda823b8fa9aac74c862605a1c60961..0000000000000000000000000000000000000000
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_2pimofs7erzvi.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:70c3d5adb831c3fa4f7fabc1490a040fe95a2b30f7fc08baeda6b15ea5d30a68
-size 84165640
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_z6j3gzsycn542.abi3.so b/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_z6j3gzsycn542.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..3103c40b5f3ec29b903b640ebc9828b0b4663781
--- /dev/null
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/_moe_z6j3gzsycn542.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9664c7b8a4e935582354443bebc5557041cac1d35b4b483abe73b4559d7c468c
+size 85827696
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/_ops.py b/build/torch25-cxx11-cu118-x86_64-linux/moe/_ops.py
index 1ad63e66a75ad1807b603125bdf8c9d4e2719648..885954104d4adf106abc62466a24346d13929950 100644
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_2pimofs7erzvi
-ops = torch.ops._moe_2pimofs7erzvi
+from . import _moe_z6j3gzsycn542
+ops = torch.ops._moe_z6j3gzsycn542
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_2pimofs7erzvi::{op_name}"
\ No newline at end of file
+    return f"_moe_z6j3gzsycn542::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/fp8.py b/build/torch25-cxx11-cu118-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx11-cu118-x86_64-linux/moe/platforms.py b/build/torch25-cxx11-cu118-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx11-cu118-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx11-cu118-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_pqwfgssq5enn2.abi3.so b/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_pqwfgssq5enn2.abi3.so
deleted file mode 100755
index 3e7d819a3fca7aed8cbe13c89599a88d3a3698b7..0000000000000000000000000000000000000000
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_pqwfgssq5enn2.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:11fd7f53b6268c0f8eeae2b752e190880de6ec16733878a8aa6b9073da2c946f
-size 84364536
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_tuji4gj3mmhfo.abi3.so b/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_tuji4gj3mmhfo.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..7163928e79c22d01198c756f24b01ab3bb90d77d
--- /dev/null
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/_moe_tuji4gj3mmhfo.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7848d33b838158269ee403fbd068b92fae716bfc27a22f393935247b9ad58848
+size 86034528
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/_ops.py b/build/torch25-cxx11-cu121-x86_64-linux/moe/_ops.py
index 30142c0ecdb4cf52ac4a8ec48443598126de4fa8..f7a5281ff3a1037b4f4bdd6367522eef4c3fbe6a 100644
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_pqwfgssq5enn2
-ops = torch.ops._moe_pqwfgssq5enn2
+from . import _moe_tuji4gj3mmhfo
+ops = torch.ops._moe_tuji4gj3mmhfo
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_pqwfgssq5enn2::{op_name}"
\ No newline at end of file
+    return f"_moe_tuji4gj3mmhfo::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/fp8.py b/build/torch25-cxx11-cu121-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx11-cu121-x86_64-linux/moe/platforms.py b/build/torch25-cxx11-cu121-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx11-cu121-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx11-cu121-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_lwzoz7knnxf4i.abi3.so b/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_lwzoz7knnxf4i.abi3.so
deleted file mode 100755
index 3b77403a421d64162163da08305b5ff13a99122e..0000000000000000000000000000000000000000
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_lwzoz7knnxf4i.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:db858df36eb173a729f2c5a99936eb0a75b92cfd795ed9080e0b05c231ed969a
-size 84063160
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_pss5doo675cd4.abi3.so b/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_pss5doo675cd4.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..3e9ddaa6894923fa39c435d2a29c6b7b042c6bd9
--- /dev/null
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/_moe_pss5doo675cd4.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:041c922d7e435dbc7ca974c331455f02ed43ecd4adcd859dd8ee593cfea676e3
+size 85733000
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/_ops.py b/build/torch25-cxx11-cu124-x86_64-linux/moe/_ops.py
index 0da7e8866438bd9519be69e87b126203f9f2e077..8d365e2f733029542609537d8d2b09fd41d2acde 100644
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_lwzoz7knnxf4i
-ops = torch.ops._moe_lwzoz7knnxf4i
+from . import _moe_pss5doo675cd4
+ops = torch.ops._moe_pss5doo675cd4
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_lwzoz7knnxf4i::{op_name}"
\ No newline at end of file
+    return f"_moe_pss5doo675cd4::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/fp8.py b/build/torch25-cxx11-cu124-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx11-cu124-x86_64-linux/moe/platforms.py b/build/torch25-cxx11-cu124-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx11-cu124-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx11-cu124-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_5uyw6qhdybj5e.abi3.so b/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_5uyw6qhdybj5e.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..e103cda878e42a7bcc8aafbda24c89bc8a2c29d1
--- /dev/null
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_5uyw6qhdybj5e.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acfcb8be6199c8e08519a1db8ec8122f7ec69a96c798d9c26e681469ba326782
+size 85815472
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_uhyif3wslpwak.abi3.so b/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_uhyif3wslpwak.abi3.so
deleted file mode 100755
index 8b5385ff774209f6abf2138700f0e4ec54145763..0000000000000000000000000000000000000000
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/_moe_uhyif3wslpwak.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ba422099b10d7e4972bb85371663a2e9765ae76cfa33c49022a34512f63e6be9
-size 84157888
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/_ops.py b/build/torch25-cxx98-cu118-x86_64-linux/moe/_ops.py
index 76014aa745c27c970ef03448fbcc74ff0c5079a5..10dfd2ce84a2e4dad3c6c20c3dd20d240f07e908 100644
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_uhyif3wslpwak
-ops = torch.ops._moe_uhyif3wslpwak
+from . import _moe_5uyw6qhdybj5e
+ops = torch.ops._moe_5uyw6qhdybj5e
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_uhyif3wslpwak::{op_name}"
\ No newline at end of file
+    return f"_moe_5uyw6qhdybj5e::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/fp8.py b/build/torch25-cxx98-cu118-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx98-cu118-x86_64-linux/moe/platforms.py b/build/torch25-cxx98-cu118-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx98-cu118-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx98-cu118-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_tj3osoay2niyk.abi3.so b/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_tj3osoay2niyk.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..6499dc1a1f242ba3c9ca65b0f74069c2c03d4a24
--- /dev/null
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_tj3osoay2niyk.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55b0eed6d5e4f8ef44d2f5baea4466cc633ae561aefd48dc54d648b9dc4742f3
+size 86026776
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_xsk7dxl7fy4pk.abi3.so b/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_xsk7dxl7fy4pk.abi3.so
deleted file mode 100755
index 80dbdcfbbb4e6aa9476af534939a44573fe932e7..0000000000000000000000000000000000000000
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/_moe_xsk7dxl7fy4pk.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:53bd3b3d77a869ea6325993ff091433f370925006947f7a8218c02c6b24fddf9
-size 84360992
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/_ops.py b/build/torch25-cxx98-cu121-x86_64-linux/moe/_ops.py
index a442144d7fc16d975ba7172458a924c459f9a007..3ed9088bee211a60b447dc35e160f5ea917d5d21 100644
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_xsk7dxl7fy4pk
-ops = torch.ops._moe_xsk7dxl7fy4pk
+from . import _moe_tj3osoay2niyk
+ops = torch.ops._moe_tj3osoay2niyk
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_xsk7dxl7fy4pk::{op_name}"
\ No newline at end of file
+    return f"_moe_tj3osoay2niyk::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/fp8.py b/build/torch25-cxx98-cu121-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx98-cu121-x86_64-linux/moe/platforms.py b/build/torch25-cxx98-cu121-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx98-cu121-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx98-cu121-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_b25pgchg5o5pa.abi3.so b/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_b25pgchg5o5pa.abi3.so
deleted file mode 100755
index 89829502267352b4ec3937e3f1fb713549bae030..0000000000000000000000000000000000000000
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_b25pgchg5o5pa.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1b5e6a3b584873f4b48185c810e8cc1045b000e45269f2490a2e2fc3a45e144b
-size 84059584
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_phlujktdbqekw.abi3.so b/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_phlujktdbqekw.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..2aee57726907c3409a2d9f397ba386ed923b71ca
--- /dev/null
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/_moe_phlujktdbqekw.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c3b1cc57c3f73b7c43aec3aa6c0673bc8e24827a0338ef8beeb431392e9ac3e
+size 85733416
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/_ops.py b/build/torch25-cxx98-cu124-x86_64-linux/moe/_ops.py
index d89a5ec6f9a1e64b3b0579eaab528abc53380e84..43cbbf7d2c681debf3f2a261091cf2e88151e1ec 100644
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/_ops.py
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_b25pgchg5o5pa
-ops = torch.ops._moe_b25pgchg5o5pa
+from . import _moe_phlujktdbqekw
+ops = torch.ops._moe_phlujktdbqekw
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_b25pgchg5o5pa::{op_name}"
\ No newline at end of file
+    return f"_moe_phlujktdbqekw::{op_name}"
\ No newline at end of file
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/fp8.py b/build/torch25-cxx98-cu124-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/fp8.py
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py b/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_moe.py b/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_moe.py
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch25-cxx98-cu124-x86_64-linux/moe/platforms.py b/build/torch25-cxx98-cu124-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch25-cxx98-cu124-x86_64-linux/moe/platforms.py
+++ b/build/torch25-cxx98-cu124-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_ooomuvan6f6yy.abi3.so b/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_ooomuvan6f6yy.abi3.so
deleted file mode 100755
index 4119dc3242f48a2077fbc79c063dd6ec75f825c6..0000000000000000000000000000000000000000
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_ooomuvan6f6yy.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1de7247bc801effbb2c8698bb47eddb97a57baeea9fb7bb05f70f42d0db0ab7f
-size 84165848
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_zlz7rpd2goyn2.abi3.so b/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_zlz7rpd2goyn2.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..164068fd1a85272b8ab3bdb3b6d57ebc24ba1453
--- /dev/null
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/_moe_zlz7rpd2goyn2.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:658fb6f129cf6ba0ea172ccfd1f115c0a03e5574122456ab9ecd35122908369a
+size 85823776
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/_ops.py b/build/torch26-cxx11-cu118-x86_64-linux/moe/_ops.py
index f6db35ecd7d5daf9bf3a6c33c791ee715b8e4e18..119443358d02a603d12a22e7bf95d416094ba9d5 100644
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_ooomuvan6f6yy
-ops = torch.ops._moe_ooomuvan6f6yy
+from . import _moe_zlz7rpd2goyn2
+ops = torch.ops._moe_zlz7rpd2goyn2
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_ooomuvan6f6yy::{op_name}"
\ No newline at end of file
+    return f"_moe_zlz7rpd2goyn2::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/fp8.py b/build/torch26-cxx11-cu118-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx11-cu118-x86_64-linux/moe/platforms.py b/build/torch26-cxx11-cu118-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx11-cu118-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx11-cu118-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_h5rxhm5fum47w.abi3.so b/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_h5rxhm5fum47w.abi3.so
deleted file mode 100755
index 45a7691711dfa415ea34138b61845daf56bef16b..0000000000000000000000000000000000000000
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_h5rxhm5fum47w.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:82358e87c49352e80bf23b7cbb9e52ed655be254b7da552ebdaa5af172a8625f
-size 84063432
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_wua27hyvpwmli.abi3.so b/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_wua27hyvpwmli.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..7ec24da2cc15da13e59aa016350f01f56535301d
--- /dev/null
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/_moe_wua27hyvpwmli.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3f7f1fa2f76004fba0e0d4eb8cbc3e35a7182538c83261f4a01a8e7401bfa81
+size 85737400
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/_ops.py b/build/torch26-cxx11-cu124-x86_64-linux/moe/_ops.py
index 712ac262bd80bdac6b8145bb28fd73a2634ae78e..8aa91e54dfdf049e1ec0b9407b3b97e6ad6a4369 100644
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_h5rxhm5fum47w
-ops = torch.ops._moe_h5rxhm5fum47w
+from . import _moe_wua27hyvpwmli
+ops = torch.ops._moe_wua27hyvpwmli
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_h5rxhm5fum47w::{op_name}"
\ No newline at end of file
+    return f"_moe_wua27hyvpwmli::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/fp8.py b/build/torch26-cxx11-cu124-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx11-cu124-x86_64-linux/moe/platforms.py b/build/torch26-cxx11-cu124-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx11-cu124-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx11-cu124-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_3z4bgea4nke26.abi3.so b/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_3z4bgea4nke26.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..5a9c623d988a1fdd82e68b6549199d498f455e6b
--- /dev/null
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_3z4bgea4nke26.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1c7676c910bb702d77a6dbf1653d9f17876924502bbbfc6661b85b8eaa0969d
+size 86192320
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_rybwc37z6ntl4.abi3.so b/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_rybwc37z6ntl4.abi3.so
deleted file mode 100755
index 48d2d0c7666b0fd0fdecd4a2b899e44884424add..0000000000000000000000000000000000000000
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/_moe_rybwc37z6ntl4.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:51ef4963f489e7bc0efc8312474a4194ae6b3039c2463b9a403b757be8ac83ee
-size 84500688
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/_ops.py b/build/torch26-cxx11-cu126-x86_64-linux/moe/_ops.py
index 6a940acbf5b493299d5fe99700de3c1e4fe6776a..9e21d2214643c15a2c652dab6f5afaa3ecc02748 100644
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_rybwc37z6ntl4
-ops = torch.ops._moe_rybwc37z6ntl4
+from . import _moe_3z4bgea4nke26
+ops = torch.ops._moe_3z4bgea4nke26
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_rybwc37z6ntl4::{op_name}"
\ No newline at end of file
+    return f"_moe_3z4bgea4nke26::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/fp8.py b/build/torch26-cxx11-cu126-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx11-cu126-x86_64-linux/moe/platforms.py b/build/torch26-cxx11-cu126-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx11-cu126-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx11-cu126-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_6pqosodmbqdcu.abi3.so b/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_6pqosodmbqdcu.abi3.so
deleted file mode 100755
index 8da38faaa997e5ea2bd72eef9674434ac0191dd3..0000000000000000000000000000000000000000
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_6pqosodmbqdcu.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f271be918718e0db8aa1a37eec3442619d98f9413b48a2a9176735a8832f04cc
-size 84158104
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_ecknt47nyrfxy.abi3.so b/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_ecknt47nyrfxy.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..92695c5f52d50d793411a6e9dd993ac5fd7f58a3
--- /dev/null
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/_moe_ecknt47nyrfxy.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f81a1d74110cc6f4c75d299b0bfa42b4789fd658d167d78c8786c0e10b08d1e
+size 85820040
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/_ops.py b/build/torch26-cxx98-cu118-x86_64-linux/moe/_ops.py
index 59e1ec3526346abfdb0c396dd1929f12d9330759..1a554ec166b5d7bd7fbe5923b07015fe9524fafa 100644
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_6pqosodmbqdcu
-ops = torch.ops._moe_6pqosodmbqdcu
+from . import _moe_ecknt47nyrfxy
+ops = torch.ops._moe_ecknt47nyrfxy
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_6pqosodmbqdcu::{op_name}"
\ No newline at end of file
+    return f"_moe_ecknt47nyrfxy::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/fp8.py b/build/torch26-cxx98-cu118-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx98-cu118-x86_64-linux/moe/platforms.py b/build/torch26-cxx98-cu118-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx98-cu118-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx98-cu118-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_xxftygesccbmy.abi3.so b/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_xxftygesccbmy.abi3.so
deleted file mode 100755
index ae9ffd81527c442d8c72a63c6209c9d1823a138a..0000000000000000000000000000000000000000
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_xxftygesccbmy.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d92104e581662916f0841da068fbc3e87ff230da2afebc6c725462a8be5b6f7c
-size 84059808
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_zirytomtyvq4i.abi3.so b/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_zirytomtyvq4i.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..9ec1b893f30459b17f911a1b7f6ed0e7d79e2811
--- /dev/null
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/_moe_zirytomtyvq4i.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:333baa2d499c511f0f2f2b81a21eee2e21e8bed8f45311d222da690a59b7ad4e
+size 85725672
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/_ops.py b/build/torch26-cxx98-cu124-x86_64-linux/moe/_ops.py
index 74e9536845edee9803eb008f9a37af3bf7096dd6..1258cbfe4e9f2c65476d403bf31763d0704497aa 100644
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_xxftygesccbmy
-ops = torch.ops._moe_xxftygesccbmy
+from . import _moe_zirytomtyvq4i
+ops = torch.ops._moe_zirytomtyvq4i
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_xxftygesccbmy::{op_name}"
\ No newline at end of file
+    return f"_moe_zirytomtyvq4i::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/fp8.py b/build/torch26-cxx98-cu124-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx98-cu124-x86_64-linux/moe/platforms.py b/build/torch26-cxx98-cu124-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx98-cu124-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx98-cu124-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_cvfkca6s5srfc.abi3.so b/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_cvfkca6s5srfc.abi3.so
new file mode 100755
index 0000000000000000000000000000000000000000..a629b6120599fd93b5e29b599a7ffb4f3aa95d57
--- /dev/null
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_cvfkca6s5srfc.abi3.so
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:169328fb13eb33abdbde9044fba8e9bf958041ca5217ce1f6dee29a5eca62dff
+size 86184688
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_jgwmm3wsss76o.abi3.so b/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_jgwmm3wsss76o.abi3.so
deleted file mode 100755
index 21e654d691e00ae22a3f8dd16bef2577e19474a2..0000000000000000000000000000000000000000
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/_moe_jgwmm3wsss76o.abi3.so
+++ /dev/null
@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7793406ff253d34fd41230d824563acb54cd6ad5c679b1d2b35b8eb590ba0f35
-size 84492968
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/_ops.py b/build/torch26-cxx98-cu126-x86_64-linux/moe/_ops.py
index e5806594a4f7fb721e177f37626ff90bdde1d9c6..d3619818f58d76f3314abecde96052b5e82a7579 100644
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/_ops.py
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/_ops.py
@@ -1,9 +1,9 @@
 import torch
-from . import _moe_jgwmm3wsss76o
-ops = torch.ops._moe_jgwmm3wsss76o
+from . import _moe_cvfkca6s5srfc
+ops = torch.ops._moe_cvfkca6s5srfc
 
 def add_op_namespace_prefix(op_name: str):
     """
     Prefix op by namespace.
     """
-    return f"_moe_jgwmm3wsss76o::{op_name}"
\ No newline at end of file
+    return f"_moe_cvfkca6s5srfc::{op_name}"
\ No newline at end of file
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/fp8.py b/build/torch26-cxx98-cu126-x86_64-linux/moe/fp8.py
index 4f790c4b88d9c393bb31da22d1c32acd375bc010..cfc6650d9d1a2062cb7bf1d08434f9f9c2e6e5ba 100644
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/fp8.py
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/fp8.py
@@ -1,6 +1,11 @@
+from typing import Tuple, Optional, Union
+
 import torch
+import triton
+import triton.language as tl
 
-from typing import Tuple, Optional, Union
+
+from ._ops import ops
 
 
 def is_hip() -> bool:
@@ -49,15 +54,179 @@ def scaled_fp8_quant(
     if scale is None:
         if use_per_token_if_dynamic:
             scale = torch.empty((shape[0], 1), device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
-                output, input, scale, scale_ub
-            )
+            ops.dynamic_per_token_scaled_fp8_quant(output, input, scale, scale_ub)
         else:
             scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+            ops.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         # num_token_padding not implemented for this case
         assert scale.numel() == 1 or num_token_padding is None
-        torch.ops._C.static_scaled_fp8_quant(output, input, scale)
+        ops.static_scaled_fp8_quant(output, input, scale)
 
     return output, scale
+
+
+@triton.jit
+def _per_token_group_quant_fp8(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+    y_s_ptr += g_id
+
+    cols = tl.arange(0, BLOCK)  # N <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+@triton.jit
+def _per_token_group_quant_fp8_colmajor(
+    # Pointers to inputs and output
+    y_ptr,
+    y_q_ptr,
+    y_s_ptr,
+    group_size,
+    # Num columns of y
+    y_num_columns,
+    # Stride from one column to the next of y_s
+    y_s_col_stride,
+    # Avoid to divide zero
+    eps,
+    # Information for float8
+    fp8_min,
+    fp8_max,
+    # Meta-parameters
+    BLOCK: tl.constexpr,
+):
+    """A Triton-accelerated function to perform per-token-group
+    quantization on a tensor.
+    This function converts the tensor values into float8 values.
+    """
+    # Map the program id to the row of X and Y it should compute.
+    g_id = tl.program_id(0)
+    y_ptr += g_id * group_size
+    y_q_ptr += g_id * group_size
+
+    # Convert g_id the flattened block coordinate to 2D so we can index
+    # into the output y_scales matrix
+    blocks_per_row = y_num_columns // group_size
+    scale_col = g_id % blocks_per_row
+    scale_row = g_id // blocks_per_row
+    y_s_ptr += scale_col * y_s_col_stride + scale_row
+
+    cols = tl.arange(0, BLOCK)  # group_size <= BLOCK
+    mask = cols < group_size
+
+    y = tl.load(y_ptr + cols, mask=mask, other=0.0).to(tl.float32)
+    # Quant
+    _absmax = tl.maximum(tl.max(tl.abs(y)), eps)
+    y_s = _absmax / fp8_max
+    y_q = tl.clamp(y / y_s, fp8_min, fp8_max).to(y_q_ptr.dtype.element_ty)
+
+    tl.store(y_q_ptr + cols, y_q, mask=mask)
+    tl.store(y_s_ptr, y_s)
+
+
+def per_token_group_quant_fp8(
+    x: torch.Tensor,
+    group_size: int,
+    eps: float = 1e-10,
+    dtype: Optional[torch.dtype] = None,
+    column_major_scales: bool = False,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """Function to perform per-token-group quantization on an input tensor `x`.
+    It converts the tensor values into signed float8 values and returns the
+    quantized tensor along with the scaling factor used for quantization.
+    Args:
+        x: The input tensor with ndim >= 2.
+        group_size: The group size used for quantization.
+        eps: The minimum to avoid dividing zero.
+        dtype: The dype of output tensor. Note that only `torch.float8_e4m3fn`
+        is supported for now.
+    Returns:
+        Tuple[torch.Tensor, torch.Tensor]: The quantized tensor and the
+        scaling factor for quantization.
+    """
+    if dtype is None:
+        dtype = (
+            torch.float8_e4m3fnuz if current_platform.is_rocm() else torch.float8_e4m3fn
+        )
+    assert x.shape[-1] % group_size == 0, (
+        f"the last dimension of `x` {x.shape[-1]} must be divisible "
+        f"by `group_size` {group_size}"
+    )
+    assert x.is_contiguous(), "`x` must be contiguous"
+
+    finfo = torch.finfo(dtype)
+    fp8_min = finfo.min
+    fp8_max = finfo.max
+
+    x_q = torch.empty_like(x, device=x.device, dtype=dtype)
+    M = x.numel() // group_size
+    N = group_size
+    if column_major_scales:
+        shape = (x.shape[-1] // group_size,) + x.shape[:-1]
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32).permute(-1, -2)
+    else:
+        shape = x.shape[:-1] + (x.shape[-1] // group_size,)
+        x_s = torch.empty(shape, device=x.device, dtype=torch.float32)
+
+    BLOCK = triton.next_power_of_2(N)
+    # heuristics for number of warps
+    num_warps = min(max(BLOCK // 256, 1), 8)
+    num_stages = 1
+    if column_major_scales:
+        _per_token_group_quant_fp8_colmajor[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            x.shape[1],
+            x_s.stride(1),
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+    else:
+        _per_token_group_quant_fp8[(M,)](
+            x,
+            x_q,
+            x_s,
+            group_size,
+            eps,
+            fp8_min=fp8_min,
+            fp8_max=fp8_max,
+            BLOCK=BLOCK,
+            num_warps=num_warps,
+            num_stages=num_stages,
+        )
+
+    return x_q, x_s
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_marlin_moe.py b/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_marlin_moe.py
index 6655bf13b910a7fcd64102143c2d630fb8f7f224..a140923049bc34ab82565407240ccea9969910cc 100644
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_marlin_moe.py
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_marlin_moe.py
@@ -40,7 +40,6 @@ def single_marlin_moe(
     g_idx: Optional[torch.Tensor] = None,
     sort_indices: Optional[torch.Tensor] = None,
     w_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -61,8 +60,6 @@ def single_marlin_moe(
     - topk (int): The number of top-k experts to select.
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - w_zeros (Optional[torch.Tensor]): Optional zero points to be used for w.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_bits (bool): The number of bits in expert weights quantization.
 
     Returns:
@@ -90,7 +87,6 @@ def single_marlin_moe(
         w.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -154,6 +150,25 @@ def single_marlin_moe(
     return torch.sum(intermediate_cache.view(*intermediate_cache.shape), dim=1)
 
 
+if hasattr(ops, "single_marlin_gemm_moe"):
+
+    @register_fake(add_op_namespace_prefix("single_marlin_gemm_moe"))
+    def single_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w: torch.Tensor,
+        scales: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk: int,
+        renormalize: bool,
+        g_idx: Optional[torch.Tensor] = None,
+        sort_indices: Optional[torch.Tensor] = None,
+        w_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -169,7 +184,6 @@ def fused_marlin_moe(
     sort_indices2: Optional[torch.Tensor] = None,
     w1_zeros: Optional[torch.Tensor] = None,
     w2_zeros: Optional[torch.Tensor] = None,
-    override_config: Optional[Dict[str, Any]] = None,
     num_bits: int = 8,
     is_k_full: bool = True,
 ) -> torch.Tensor:
@@ -193,8 +207,6 @@ def fused_marlin_moe(
         permutation.
     - topk_weights (torch.Tensor): Top-k weights.
     - topk_ids (torch.Tensor): Indices of topk-k elements.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - w1_zeros (Optional[torch.Tensor]): Optional zero points to be used for w1.
     - w2_zeros (Optional[torch.Tensor]): Optional zero points to be used for w2.
     - num_bits (bool): The number of bits in expert weights quantization.
@@ -248,7 +260,6 @@ def fused_marlin_moe(
         w2.shape,
         topk_ids.shape[1],
         None,
-        override_config=override_config,
         is_marlin=True,
     )
     config = get_config_func(M)
@@ -350,6 +361,30 @@ def fused_marlin_moe(
     return torch.sum(intermediate_cache3.view(*intermediate_cache3.shape), dim=1)
 
 
+if hasattr(ops, "fused_marlin_moe"):
+
+    @register_fake(add_op_namespace_prefix("fused_marlin_moe"))
+    def fused_marlin_moe_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        w1_scale: torch.Tensor,
+        w2_scale: torch.Tensor,
+        gating_output: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        g_idx1: Optional[torch.Tensor] = None,
+        g_idx2: Optional[torch.Tensor] = None,
+        sort_indices1: Optional[torch.Tensor] = None,
+        sort_indices2: Optional[torch.Tensor] = None,
+        w1_zeros: Optional[torch.Tensor] = None,
+        w2_zeros: Optional[torch.Tensor] = None,
+        num_bits: int = 8,
+        is_k_full: bool = True,
+    ) -> torch.Tensor:
+        return torch.empty_like(hidden_states)
+
+
 if hasattr(ops, "marlin_gemm_moe"):
 
     @register_fake(add_op_namespace_prefix("marlin_gemm_moe"))
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_moe.py b/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_moe.py
index 49a09b7eca6bac8b0907ce11395ae5198989d531..a48ce4926942bbc5fc7f53c95173484add243fb0 100644
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_moe.py
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/fused_moe.py
@@ -1,21 +1,242 @@
+# SPDX-License-Identifier: Apache-2.0
 """Fused MoE kernel."""
 
 import functools
 import json
+import logging
 import os
-from typing import Any, Callable, Dict, Optional, Tuple
+from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import torch
 import triton
 import triton.language as tl
 
+
 from ._ops import ops
-from .fp8 import scaled_fp8_quant
+from .fp8 import per_token_group_quant_fp8, scaled_fp8_quant
 from .platforms import current_platform
 
+logger = logging.getLogger(__name__)
+
+
 VLLM_FUSED_MOE_CHUNK_SIZE = int(os.getenv("VLLM_FUSED_MOE_CHUNK_SIZE", "32768"))
 
 
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    b_scale_ptr,
+    b_zp_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N: tl.constexpr,
+    K: tl.constexpr,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    stride_bze,
+    stride_bzk,
+    stride_bzn,
+    block_k_diviable: tl.constexpr,
+    group_size: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    has_zp: tl.constexpr,
+    use_int4_w4a16: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (
+        offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
+    )
+
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
+
+    if use_int4_w4a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + (offs_k[:, None] // 2) * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = (
+            b_ptr
+            + off_experts * stride_be
+            + offs_k[:, None] * stride_bk
+            + offs_bn[None, :] * stride_bn
+        )
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(
+            a_ptrs,
+            mask=token_mask[:, None] & (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+            other=0.0,
+        )
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = (
+            b_scale_ptr
+            + off_experts * stride_bse
+            + offs_bn[None, :] * stride_bsn
+            + ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        )
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + (offs_bn[None, :] // 2) * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = (b_zp >> b_zp_shifter) & 0xF
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = (
+                b_zp_ptr
+                + off_experts * stride_bze
+                + offs_bn[None, :] * stride_bzn
+                + offs_k_true * stride_bzk
+            )
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token, mask=token_mask, other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
 @triton.jit
 def fused_moe_kernel(
     # Pointers to matrices
@@ -44,8 +265,14 @@ def fused_moe_kernel(
     stride_bn,
     stride_cm,
     stride_cn,
+    stride_asm,
+    stride_ask,
     stride_bse,
+    stride_bsk,
     stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
     # Meta-parameters
     BLOCK_SIZE_M: tl.constexpr,
     BLOCK_SIZE_N: tl.constexpr,
@@ -105,17 +332,17 @@ def fused_moe_kernel(
     num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
     if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
         return
-    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
     offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
     token_mask = offs_token < num_valid_tokens
 
-    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
     offs_k = tl.arange(0, BLOCK_SIZE_K)
     a_ptrs = a_ptr + (
         offs_token[:, None] // top_k * stride_am + offs_k[None, :] * stride_ak
     )
 
-    off_experts = tl.load(expert_ids_ptr + pid_m)
+    off_experts = tl.load(expert_ids_ptr + pid_m).to(tl.int64)
     b_ptrs = (
         b_ptr
         + off_experts * stride_be
@@ -128,8 +355,15 @@ def fused_moe_kernel(
         b_scale = tl.load(b_scale_ptrs)
 
     if use_fp8_w8a8:
-        a_scale = tl.load(a_scale_ptr)
-        b_scale = tl.load(b_scale_ptr + off_experts)
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (
+                b_scale_ptr + off_experts * stride_bse + offs_bsn * stride_bsn
+            )
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
 
     # -----------------------------------------------------------
     # Iterate to compute a block of the C matrix.
@@ -151,7 +385,17 @@ def fused_moe_kernel(
         if use_int8_w8a16:
             accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
         elif use_fp8_w8a8:
-            accumulator = tl.dot(a, b, acc=accumulator)
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(
+                    a_scale_ptrs + offs_ks * stride_ask, mask=token_mask, other=0.0
+                )
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:, None] * b_scale[None, :]
+            else:
+                accumulator = tl.dot(a, b, acc=accumulator)
         else:
             accumulator += tl.dot(a, b)
         # Advance the ptrs to the next K block.
@@ -164,7 +408,10 @@ def fused_moe_kernel(
     if use_int8_w8a16:
         accumulator = (accumulator * b_scale).to(compute_type)
     elif use_fp8_w8a8:
-        accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
     else:
         accumulator = accumulator.to(compute_type)
     # -----------------------------------------------------------
@@ -175,6 +422,141 @@ def fused_moe_kernel(
     tl.store(c_ptrs, accumulator, mask=c_mask)
 
 
+def ceil_div(a, b):
+    return (a + b - 1) // b
+
+
+@triton.jit
+def moe_align_block_size_stage1(
+    topk_ids_ptr,
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    start_idx = pid * tokens_per_thread
+
+    off_c = (pid + 1) * num_experts
+
+    for i in range(tokens_per_thread):
+        if start_idx + i < numel:
+            idx = tl.load(topk_ids_ptr + start_idx + i)
+            token_cnt = tl.load(tokens_cnts_ptr + off_c + idx)
+            tl.store(tokens_cnts_ptr + off_c + idx, token_cnt + 1)
+
+
+@triton.jit
+def moe_align_block_size_stage2(
+    tokens_cnts_ptr,
+    num_experts: tl.constexpr,
+):
+    pid = tl.program_id(0)
+
+    last_cnt = 0
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + i * num_experts + pid)
+        last_cnt = last_cnt + token_cnt
+        tl.store(tokens_cnts_ptr + i * num_experts + pid, last_cnt)
+
+
+@triton.jit
+def moe_align_block_size_stage3(
+    total_tokens_post_pad_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+):
+    last_cumsum = 0
+    off_cnt = num_experts * num_experts
+    for i in range(1, num_experts + 1):
+        token_cnt = tl.load(tokens_cnts_ptr + off_cnt + i - 1)
+        last_cumsum = last_cumsum + tl.cdiv(token_cnt, block_size) * block_size
+        tl.store(cumsum_ptr + i, last_cumsum)
+    tl.store(total_tokens_post_pad_ptr, last_cumsum)
+
+
+@triton.jit
+def moe_align_block_size_stage4(
+    topk_ids_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    tokens_cnts_ptr,
+    cumsum_ptr,
+    num_experts: tl.constexpr,
+    block_size: tl.constexpr,
+    numel: tl.constexpr,
+    tokens_per_thread: tl.constexpr,
+):
+    pid = tl.program_id(0)
+    start_idx = tl.load(cumsum_ptr + pid)
+    end_idx = tl.load(cumsum_ptr + pid + 1)
+
+    for i in range(start_idx, end_idx, block_size):
+        tl.store(expert_ids_ptr + i // block_size, pid)
+
+    start_idx = pid * tokens_per_thread
+    off_t = pid * num_experts
+
+    for i in range(start_idx, tl.minimum(start_idx + tokens_per_thread, numel)):
+        expert_id = tl.load(topk_ids_ptr + i)
+        token_cnt = tl.load(tokens_cnts_ptr + off_t + expert_id)
+        rank_post_pad = token_cnt + tl.load(cumsum_ptr + expert_id)
+        tl.store(sorted_token_ids_ptr + rank_post_pad, i)
+        tl.store(tokens_cnts_ptr + off_t + expert_id, token_cnt + 1)
+
+
+# Triton implementation based on:
+# https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0
+def moe_align_block_size_triton(
+    topk_ids: torch.Tensor,
+    num_experts: int,
+    block_size: int,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_pad: torch.Tensor,
+) -> None:
+    numel = topk_ids.numel()
+    grid = (num_experts,)
+    tokens_cnts = torch.zeros(
+        (num_experts + 1, num_experts), dtype=torch.int32, device=topk_ids.device
+    )
+    cumsum = torch.zeros((num_experts + 1,), dtype=torch.int32, device=topk_ids.device)
+    tokens_per_thread = ceil_div(numel, num_experts)
+
+    moe_align_block_size_stage1[grid](
+        topk_ids,
+        tokens_cnts,
+        num_experts,
+        numel,
+        tokens_per_thread,
+    )
+    moe_align_block_size_stage2[grid](
+        tokens_cnts,
+        num_experts,
+    )
+    moe_align_block_size_stage3[(1,)](
+        num_tokens_post_pad,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+    )
+    moe_align_block_size_stage4[grid](
+        topk_ids,
+        sorted_token_ids,
+        expert_ids,
+        tokens_cnts,
+        cumsum,
+        num_experts,
+        block_size,
+        numel,
+        tokens_per_thread,
+    )
+
+
 def moe_align_block_size(
     topk_ids: torch.Tensor, block_size: int, num_experts: int
 ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
@@ -225,9 +607,34 @@ def moe_align_block_size(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
     )
     num_tokens_post_pad = torch.empty((1), dtype=torch.int32, device=topk_ids.device)
-    ops.moe_align_block_size(
-        topk_ids, num_experts, block_size, sorted_ids, expert_ids, num_tokens_post_pad
-    )
+    if num_experts >= 224:
+        if VLLM_ENABLE_MOE_ALIGN_BLOCK_SIZE_TRITON:
+            moe_align_block_size_triton(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+        else:
+            ops.sgl_moe_align_block_size(
+                topk_ids,
+                num_experts,
+                block_size,
+                sorted_ids,
+                expert_ids,
+                num_tokens_post_pad,
+            )
+    else:
+        ops.moe_align_block_size(
+            topk_ids,
+            num_experts,
+            block_size,
+            sorted_ids,
+            expert_ids,
+            num_tokens_post_pad,
+        )
     return sorted_ids, expert_ids, num_tokens_post_pad
 
 
@@ -237,6 +644,7 @@ def invoke_fused_moe_kernel(
     C: torch.Tensor,
     A_scale: Optional[torch.Tensor],
     B_scale: Optional[torch.Tensor],
+    B_zp: Optional[torch.Tensor],
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     sorted_token_ids: torch.Tensor,
@@ -248,64 +656,147 @@ def invoke_fused_moe_kernel(
     compute_type: tl.dtype,
     use_fp8_w8a8: bool,
     use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> None:
     assert topk_weights.stride(1) == 1
     assert sorted_token_ids.stride(0) == 1
 
     if use_fp8_w8a8:
-        A, A_scale = scaled_fp8_quant(A, A_scale)
         assert B_scale is not None
-    elif use_int8_w8a16:
+        if block_shape is None:
+            A, A_scale = scaled_fp8_quant(A, A_scale)
+        else:
+            assert len(block_shape) == 2
+            block_n, block_k = block_shape[0], block_shape[1]
+            A, A_scale = per_token_group_quant_fp8(A, block_k)
+            assert triton.cdiv(A.shape[-1], block_k) == A_scale.shape[-1]
+            assert triton.cdiv(B.shape[-2], block_n) == B_scale.shape[-2]
+            assert triton.cdiv(B.shape[-1], block_k) == B_scale.shape[-1]
+    elif use_int8_w8a16 or use_int4_w4a16:
         assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
     else:
         assert A_scale is None
         assert B_scale is None
 
+    EM = sorted_token_ids.shape[0]
+    if A.shape[0] < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.shape[0], A.shape[0] * top_k * config["BLOCK_SIZE_M"])
     grid = lambda META: (
-        triton.cdiv(sorted_token_ids.shape[0], META["BLOCK_SIZE_M"])
+        triton.cdiv(EM, META["BLOCK_SIZE_M"])
         * triton.cdiv(B.shape[1], META["BLOCK_SIZE_N"]),
     )
 
-    fused_moe_kernel[grid](
-        A,
-        B,
-        C,
-        A_scale,
-        B_scale,
-        topk_weights,
-        sorted_token_ids,
-        expert_ids,
-        num_tokens_post_padded,
-        B.shape[1],
-        B.shape[2],
-        sorted_token_ids.shape[0],
-        topk_ids.numel(),
-        A.stride(0),
-        A.stride(1),
-        B.stride(0),
-        B.stride(2),
-        B.stride(1),
-        C.stride(1),
-        C.stride(2),
-        B_scale.stride(0) if B_scale is not None and use_int8_w8a16 else 0,
-        B_scale.stride(1) if B_scale is not None and use_int8_w8a16 else 0,
-        MUL_ROUTED_WEIGHT=mul_routed_weight,
-        top_k=top_k,
-        compute_type=compute_type,
-        use_fp8_w8a8=use_fp8_w8a8,
-        use_int8_w8a16=use_int8_w8a16,
-        **config,
-    )
+    if (
+        (use_int8_w8a16 or use_int4_w4a16)
+        and block_shape is not None
+        and block_shape[1] > 0
+    ):
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+
+        fused_moe_kernel_gptq_awq[grid](
+            A,
+            B,
+            C,
+            B_scale,
+            B_zp,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            B_scale.stride(0),
+            B_scale.stride(2),
+            B_scale.stride(1),
+            B_zp.stride(0) if B_zp is not None else 0,
+            B_zp.stride(2) if B_zp is not None else 0,
+            B_zp.stride(1) if B_zp is not None else 0,
+            block_k_diviable=A.shape[1] % config["BLOCK_SIZE_K"] == 0,
+            group_size=block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            has_zp=B_zp is not None,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
+
+    else:
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.shape[1],
+            A.shape[1],
+            EM,
+            topk_ids.numel(),
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2),
+            B.stride(1),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0) if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1) if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2) if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1) if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            **config,
+        )
 
 
-def get_config_file_name(E: int, N: int, dtype: Optional[str]) -> str:
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(
+    E: int, N: int, dtype: Optional[str], block_shape: Optional[List[int]] = None
+) -> str:
     device_name = current_platform.get_device_name().replace(" ", "_")
     dtype_selector = "" if not dtype else f",dtype={dtype}"
-    return f"E={E},N={N},device_name={device_name}{dtype_selector}.json"
+    block_shape_selector = (
+        "" if not block_shape or not all(block_shape) else f",block_shape={block_shape}"
+    )
+    return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
 
 
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
 @functools.lru_cache
-def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int, Any]]:
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+) -> Optional[Dict[int, Any]]:
     """
     Return optimized configurations for the fused MoE kernel.
 
@@ -317,18 +808,27 @@ def get_moe_configs(E: int, N: int, dtype: Optional[str]) -> Optional[Dict[int,
 
     # First look up if an optimized configuration is available in the configs
     # directory
-    json_file_name = get_config_file_name(E, N, dtype)
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape)
 
     config_file_path = os.path.join(
         os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name
     )
     if os.path.exists(config_file_path):
         with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.", config_file_path)
             # If a configuration has been found, return it
             return {int(key): val for key, val in json.load(f).items()}
 
     # If no optimized configuration is available, we will use the default
     # configuration
+    logger.warning(
+        (
+            "Using default MoE config. Performance might be sub-optimal! "
+            "Config file not found at %s"
+        ),
+        config_file_path,
+    )
     return None
 
 
@@ -340,21 +840,34 @@ def get_default_config(
     topk: int,
     dtype: Optional[str],
     is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
 ) -> Dict[str, int]:
-    config = {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 32,
-        "GROUP_SIZE_M": 8,
-    }
-    # A heuristic: fused marlin works faster with this config for small M
-    if M <= E or (is_marlin and M <= 32):
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
         config = {
-            "BLOCK_SIZE_M": 16,
-            "BLOCK_SIZE_N": 32,
-            "BLOCK_SIZE_K": 64,
-            "GROUP_SIZE_M": 1,
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3,
         }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+        # A heuristic: fused marlin works faster with this config for small M
+        if M <= E or (is_marlin and M <= 32):
+            config = {
+                "BLOCK_SIZE_M": 16,
+                "BLOCK_SIZE_N": 32,
+                "BLOCK_SIZE_K": 64,
+                "GROUP_SIZE_M": 1,
+            }
     return config
 
 
@@ -364,15 +877,21 @@ def try_get_optimal_moe_config(
     top_k: int,
     dtype: Optional[str],
     M: int,
-    override_config: Optional[Dict[str, Any]] = None,
     is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
 ):
+    # from vllm.model_executor.layers.fused_moe import get_config
+    # TODO: removed when syncing to vLLM, do we need this?
+    # override_config = get_config()
+    override_config = None
     if override_config:
         config = override_config
     else:
         # First try to load optimal config from the file
         E, _, N = w2_shape
-        configs = get_moe_configs(E, N, dtype)
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k)
 
         if configs:
             # If an optimal configuration map has been found, look up the
@@ -380,7 +899,9 @@ def try_get_optimal_moe_config(
             config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
         else:
             # Else use the default config
-            config = get_default_config(M, E, N, w1_shape[2], top_k, dtype, is_marlin)
+            config = get_default_config(
+                M, E, N, w1_shape[2], top_k, dtype, is_marlin, block_shape
+            )
     return config
 
 
@@ -416,7 +937,8 @@ def fused_topk(
     return topk_weights, topk_ids
 
 
-# This is used by the Deepseek-V2 model
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
 def grouped_topk(
     hidden_states: torch.Tensor,
     gating_output: torch.Tensor,
@@ -424,11 +946,25 @@ def grouped_topk(
     renormalize: bool,
     num_expert_group: int = 0,
     topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None,
 ):
 
     assert hidden_states.shape[0] == gating_output.shape[0], "Number of tokens mismatch"
 
-    scores = torch.softmax(gating_output, dim=-1)
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+
     num_token = scores.shape[0]
     group_scores = (
         scores.view(num_token, num_expert_group, -1).max(dim=-1).values
@@ -444,7 +980,13 @@ def grouped_topk(
         .reshape(num_token, -1)
     )  # [n, e]
     tmp_scores = scores.masked_fill(~score_mask.bool(), 0.0)  # [n, e]
-    topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)
 
     if renormalize:
         topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
@@ -454,6 +996,7 @@ def grouped_topk(
 
 def get_config_dtype_str(
     dtype: torch.dtype,
+    use_int4_w4a16: Optional[bool] = False,
     use_int8_w8a16: Optional[bool] = False,
     use_fp8_w8a8: Optional[bool] = False,
 ):
@@ -461,6 +1004,8 @@ def get_config_dtype_str(
         return "fp8_w8a8"
     elif use_int8_w8a16:
         return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w8a16"
     elif dtype == torch.float:
         # avoiding cases where kernel fails when float32 MoE
         # use fp16/bfloat16 configs
@@ -468,6 +1013,80 @@ def get_config_dtype_str(
     return None
 
 
+def inplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> None:
+    fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        True,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
+def outplace_fused_experts(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+) -> torch.Tensor:
+    return fused_experts_impl(
+        hidden_states,
+        w1,
+        w2,
+        topk_weights,
+        topk_ids,
+        False,
+        use_fp8_w8a8,
+        use_int8_w8a16,
+        use_int4_w4a16,
+        w1_scale,
+        w2_scale,
+        w1_zp,
+        w2_zp,
+        a1_scale,
+        a2_scale,
+        block_shape,
+    )
+
+
 def fused_experts(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,
@@ -475,16 +1094,80 @@ def fused_experts(
     topk_weights: torch.Tensor,
     topk_ids: torch.Tensor,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+):
+    if inplace:
+        inplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+        return hidden_states
+    else:
+        return outplace_fused_experts(
+            hidden_states,
+            w1,
+            w2,
+            topk_weights,
+            topk_ids,
+            use_fp8_w8a8,
+            use_int8_w8a16,
+            use_int4_w4a16,
+            w1_scale,
+            w2_scale,
+            w1_zp,
+            w2_zp,
+            a1_scale,
+            a2_scale,
+            block_shape,
+        )
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ):
     # Check constraints.
-    assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+    if use_int4_w4a16:
+        assert hidden_states.shape[1] // 2 == w1.shape[2], "Hidden size mismatch"
+    else:
+        assert hidden_states.shape[1] == w1.shape[2], "Hidden size mismatch"
+
     assert topk_weights.shape == topk_ids.shape, "topk shape mismatch"
     assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
     assert w1.is_contiguous(), "Expert weights1 must be contiguous"
@@ -500,6 +1183,7 @@ def fused_experts(
     config_dtype = get_config_dtype_str(
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         dtype=hidden_states.dtype,
     )
 
@@ -509,7 +1193,7 @@ def fused_experts(
         w2.shape,
         topk_ids.shape[1],
         config_dtype,
-        override_config=override_config,
+        block_shape=block_shape,
     )
 
     config = get_config_func(M)
@@ -530,7 +1214,14 @@ def fused_experts(
         dtype=hidden_states.dtype,
     )
 
-    compute_type = tl.bfloat16 if hidden_states.dtype == torch.bfloat16 else tl.float16
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
 
     if inplace:
         out_hidden_states = hidden_states
@@ -571,6 +1262,7 @@ def fused_experts(
             intermediate_cache1,
             a1_scale,
             w1_scale,
+            w1_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -582,6 +1274,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.silu_and_mul(intermediate_cache2, intermediate_cache1.view(-1, N))
@@ -592,6 +1286,7 @@ def fused_experts(
             intermediate_cache3,
             a2_scale,
             w2_scale,
+            w2_zp,
             curr_topk_weights,
             curr_topk_ids,
             sorted_token_ids,
@@ -603,6 +1298,8 @@ def fused_experts(
             compute_type=compute_type,
             use_fp8_w8a8=use_fp8_w8a8,
             use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            block_shape=block_shape,
         )
 
         ops.moe_sum(
@@ -620,17 +1317,20 @@ def fused_moe(
     topk: int,
     renormalize: bool,
     inplace: bool = False,
-    override_config: Optional[Dict[str, Any]] = None,
     use_grouped_topk: bool = False,
     num_expert_group: Optional[int] = None,
     topk_group: Optional[int] = None,
     custom_routing_function: Optional[Callable] = None,
     use_fp8_w8a8: bool = False,
     use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
     w1_scale: Optional[torch.Tensor] = None,
     w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> torch.Tensor:
     """
     This function computes a Mixture of Experts (MoE) layer using two sets of
@@ -646,20 +1346,28 @@ def fused_moe(
     - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
     - inplace (bool): If True, perform the operation in-place.
         Defaults to False.
-    - override_config (Optional[Dict[str, Any]]): Optional override
-        for the kernel configuration.
     - num_expert_group: Optional[int]: additional parameter for grouped_topk
     - topk_group: Optional[int]: additional parameter for grouped_topk
     - use_grouped_topk: If True, use grouped_topk instead of fused_topk
         note: Deepseekv2 model uses grouped_topk
     - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
         products for w1 and w2. Defaults to False.
-    - use_int8_w8a16 (bool): If True, use fp8 arithmetic to compute the inner
-        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
     - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
         w1.
     - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
         w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
 
     Returns:
     - torch.Tensor: The output tensor after applying the MoE layer.
@@ -693,11 +1401,14 @@ def fused_moe(
         topk_weights,
         topk_ids,
         inplace=inplace,
-        override_config=override_config,
         use_fp8_w8a8=use_fp8_w8a8,
         use_int8_w8a16=use_int8_w8a16,
+        use_int4_w4a16=use_int4_w4a16,
         w1_scale=w1_scale,
         w2_scale=w2_scale,
+        w1_zp=w1_zp,
+        w2_zp=w2_zp,
         a1_scale=a1_scale,
         a2_scale=a2_scale,
+        block_shape=block_shape,
     )
diff --git a/build/torch26-cxx98-cu126-x86_64-linux/moe/platforms.py b/build/torch26-cxx98-cu126-x86_64-linux/moe/platforms.py
index fb7fbbfb6c6ecdfa64901568a2c2893dd7ecae21..084120e031bd1aba9cd240941c2040a004a78c6b 100644
--- a/build/torch26-cxx98-cu126-x86_64-linux/moe/platforms.py
+++ b/build/torch26-cxx98-cu126-x86_64-linux/moe/platforms.py
@@ -1,22 +1,32 @@
-from typing import Callable, ParamSpec, TypeVar
-import os
-from functools import lru_cache, wraps
+from functools import lru_cache
 
 import torch
 
 IS_ROCM = torch.version.hip is not None
 
-class CudaPlatform:
+
+class Platform:
+    simple_compile_backend: str = "inductor"
+
+
+class CudaPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(0)
 
-class RocmPlatform:
+    def is_rocm(self):
+        return False
+
+
+class RocmPlatform(Platform):
     @classmethod
     @lru_cache(maxsize=8)
     def get_device_name(cls, device_id: int = 0) -> str:
         return torch.cuda.get_device_name(device_id)
 
+    def is_rocm(self):
+        return True
+
 
 current_platform = RocmPlatform() if IS_ROCM else CudaPlatform()