diff --git a/sgl-kernel/python/sgl_kernel/__init__.py b/sgl-kernel/python/sgl_kernel/__init__.py
index 44fe6d45c..58f06c980 100644
--- a/sgl-kernel/python/sgl_kernel/__init__.py
+++ b/sgl-kernel/python/sgl_kernel/__init__.py
@@ -34,7 +34,7 @@ from sgl_kernel.elementwise import (
     silu_and_mul,
 )
 from sgl_kernel.expert_specialization import es_fp8_blockwise_scaled_grouped_mm
-from sgl_kernel.fused_moe import fused_marlin_moe
+from sgl_kernel.fused_moe import fused_marlin_moe, moe_wna16_marlin_gemm
 from sgl_kernel.gemm import (
     awq_dequantize,
     bmm_fp8,
diff --git a/sgl-kernel/python/sgl_kernel/fused_moe.py b/sgl-kernel/python/sgl_kernel/fused_moe.py
index cc2e00fd1..8e0cea934 100644
--- a/sgl-kernel/python/sgl_kernel/fused_moe.py
+++ b/sgl-kernel/python/sgl_kernel/fused_moe.py
@@ -15,6 +15,60 @@ def get_scalar_type(num_bits: int, has_zp: bool):
         return scalar_types.uint4b8 if num_bits == 4 else scalar_types.uint8b128
 
 
+def moe_wna16_marlin_gemm(
+    a: torch.Tensor,
+    c_or_none: Optional[torch.Tensor],
+    b_q_weight: torch.Tensor,
+    b_scales: torch.Tensor,
+    b_zeros_or_none: Optional[torch.Tensor],
+    g_idx_or_none: Optional[torch.Tensor],
+    perm_or_none: Optional[torch.Tensor],
+    workspace: torch.Tensor,
+    sorted_token_ids: torch.Tensor,
+    expert_ids: torch.Tensor,
+    num_tokens_post_padded: torch.Tensor,
+    topk_weights: torch.Tensor,
+    moe_block_size: int,
+    top_k: int,
+    mul_topk_weights: bool,
+    is_ep: bool,
+    b_q_type_id: int,
+    size_m: int,
+    size_n: int,
+    size_k: int,
+    is_k_full: bool,
+    use_atomic_add: bool,
+    use_fp32_reduce: bool,
+    is_zp_float: bool,
+):
+    return torch.ops.sgl_kernel.moe_wna16_marlin_gemm.default(
+        a,
+        c_or_none,
+        b_q_weight,
+        b_scales,
+        b_zeros_or_none,
+        g_idx_or_none,
+        perm_or_none,
+        workspace,
+        sorted_token_ids,
+        expert_ids,
+        num_tokens_post_padded,
+        topk_weights,
+        moe_block_size=moe_block_size,
+        top_k=top_k,
+        mul_topk_weights=mul_topk_weights,
+        is_ep=is_ep,
+        b_q_type_id=b_q_type_id,
+        size_m=size_m,
+        size_n=size_n,
+        size_k=size_k,
+        is_k_full=is_k_full,
+        use_atomic_add=use_atomic_add,
+        use_fp32_reduce=use_fp32_reduce,
+        is_zp_float=is_zp_float,
+    )
+
+
 def fused_marlin_moe(
     hidden_states: torch.Tensor,
     w1: torch.Tensor,