Fix BadRequestError

✅ Solution

BadRequestError in vllm usually arises from invalid input provided in the API request, such as exceeding `max_tokens` or having incorrectly formatted `tool_calls`. To fix, implement thorough input validation using Pydantic models with specific data types and range checks before processing the request in vllm, and ensure the error messages clearly indicate the invalid field. For max_tokens, return HTTP 413 (Request Entity Too Large) instead of 400 when the limit is exceeded.