[Docs] Fix stale W4A16 NVFP4 default kernel in ModelOpt docs (#59135)

Signed-off-by: Rajeev Warrier <warrier.rajeev.95@gmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

docs/features/quantization/modelopt.md

@@ -33,9 +33,7 @@ following `quantization.quant_algo` values:
     `flashinfer_cutlass`, `flashinfer_cutedsl`, `flashinfer_trtllm`,
     `flashinfer_cudnn`, and `marlin`; see `KernelConfig` on the
     [Engine Arguments](../../configuration/engine_args.md) page and shown by
-    `vllm serve --help=KernelConfig`. For `W4A16_NVFP4`, `auto` currently
-    selects Marlin. Models with BF16 activations can explicitly select the
-    FlashInfer CuTe-DSL backend with `--linear-backend flashinfer_cutedsl`.
+    `vllm serve --help=KernelConfig`.
 
 !!! note
     For models quantized to MXFP8 with BF16 activations on SM100-family GPUs,

vllm/model_executor/layers/quantization/modelopt.py

@@ -2314,9 +2314,9 @@ def select_linear_kernel(
     w = spec.weight
     assert isinstance(w, QuantKey), f"resolve() must supply a weight key, got {w!r}"
     if w.dtype == FP4_DTYPE:
-        # W4A16 (activation is None) → use_a16=True defaults to Marlin *and*
-        # honors --linear-backend (matches upstream ModelOptNvFp4W4A16LinearMethod
-        # after #50273); W4A4 → use_a16=False.
+        # W4A16 (activation is None) → use_a16=True, auto prefers FlashInfer
+        # CuTe-DSL backend on SM100/103 when available and Marlin otherwise,
+        # and honors --linear-backend (#50273); W4A4 → use_a16=False.
         return init_nvfp4_linear_kernel(use_a16=spec.activation is None)
     if w.scale.dtype == MXFP8_SCALE_DTYPE:
         return init_mxfp8_linear_kernel(