What would you like to see?

We should migrate the OpenAI LLM provider to use the responses API since the responses API has better support for newer models like any O series model and gpt-5 models.