diff --git a/vllm/model_executor/layers/quantization/utils/w8a8_utils.py b/vllm/model_executor/layers/quantization/utils/w8a8_utils.py
index 4fbf75b2ff090..20100c76bd690 100644
--- a/vllm/model_executor/layers/quantization/utils/w8a8_utils.py
+++ b/vllm/model_executor/layers/quantization/utils/w8a8_utils.py
@@ -104,7 +104,7 @@ def apply_fp8_linear(
     input: torch.Tensor,
     weight: torch.Tensor,
     weight_scale: torch.Tensor,
-    input_scale: torch.Tensor,
+    input_scale: Optional[torch.Tensor] = None,
     input_scale_ub: Optional[torch.Tensor] = None,
     bias: Optional[torch.Tensor] = None,
     cutlass_fp8_supported: bool = True,
@@ -192,7 +192,7 @@ def apply_int8_linear(
     input: torch.Tensor,
     weight: torch.Tensor,
     weight_scale: torch.Tensor,
-    input_scale: torch.Tensor,
+    input_scale: Optional[torch.Tensor] = None,
     bias: Optional[torch.Tensor] = None,
 ):
     # ops.scaled_int8_quant supports both dynamic and static quant.