[Question]: Any plan to migrate OpenAI integration from chat.completions to Responses API?

Hi maintainers,

First of all, thank you for the great work on this project.

I noticed that the current OpenAI integration is still based on the `chat.completions` API. As OpenAI has introduced the new `Responses` API as the primary and future-facing interface, I would like to ask:

Is there any plan or roadmap to migrate the OpenAI model integration from `chat.completions` to the `responses` API?

From what I understand, the `Responses` API provides:

- Native multi-modal input support
- Unified structured output (JSON schema)
- Improved tool calling abstraction
- Reasoning token support
- A more future-proof unified model interface

Since `chat.completions` is still supported but no longer the recommended long-term interface, I am wondering whether this project plans to:

1. Fully migrate to `Responses`
2. Provide dual support (chat + responses)
3. Or keep `chat.completions` for compatibility reasons

If there are architectural considerations (e.g., abstraction compatibility, cross-provider consistency, or SDK limitations), I would appreciate any insights.

Thank you again for your work!