vllm

[Rust Frontend] Truncate prompts against max_model_len like Python

#62642

Merged