diff --git a/vllm/model_executor/layers/quantization/utils/marlin_utils_fp4.py b/vllm/model_executor/layers/quantization/utils/marlin_utils_fp4.py
index e4a2ab413..d6b32c4bb 100644
--- a/vllm/model_executor/layers/quantization/utils/marlin_utils_fp4.py
+++ b/vllm/model_executor/layers/quantization/utils/marlin_utils_fp4.py
@@ -37,9 +37,6 @@ def _nvfp4_compute_scale_factor(marlin_scales: torch.Tensor) -> float:
         min_val = ws_float[nonzero_mask].min()
         if min_val < 2:
             sf = (2 / min_val).log2().ceil().exp2()
-            assert (ws_float[nonzero_mask] * sf <= 448 * (2**7)).all(), (
-                "NVFP4 scale dynamic range too large for rescaling"
-            )
             return sf.item()
     return 1.0