async fn process_llm_response(response: Response) -> Result<LLMResponse>
Process a non-streaming LLM response