[BUGFIX][Qwen3.5] Hardcode `mlp.gate` as not quantizable (#35156)

Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>

vllm/model_executor/models/qwen3_next.py

@@ -252,7 +252,7 @@ class Qwen3NextSparseMoeBlock(nn.Module):
             config.hidden_size,
             config.num_experts,
             bias=False,
-            quant_config=quant_config,
+            quant_config=None,
             prefix=f"{prefix}.gate",
         )