Update vLLM description

README.md

@@ -109,6 +109,13 @@ To serve this checkpoint with [SGLang](https://github.com/sgl-project/sglang), y
 python3 -m sglang.launch_server --model nvidia/MiniMax-M2.5-NVFP4 --tensor-parallel-size 8 --quantization modelopt_fp4 --trust-remote-code --reasoning-parser minimax-append-think --tool-call-parser minimax-m2 --moe-runner-backend flashinfer_cutlass --attention-backend flashinfer
 ```
 
+To serve this checkpoint with [vLLM](https://github.com/vllm-project/vllm), you can install the [vllm nightly](https://docs.vllm.ai/en/latest/getting_started/installation/gpu/#install-the-latest-code) or start the docker `vllm/vllm-openai:nightly`, and run the sample command below: 
+
+```sh
+vllm serve nvidia/MiniMax-M2.5-NVFP4 --tensor-parallel-size 8 --trust-remote-code --reasoning-parser minimax-append-think --tool-call-parser minimax-m2  --enable-auto-tool-choice
+```
+
+
 ### Evaluation
 The accuracy benchmark results are presented in the table below:
 <table>