Feature: Support OpenResponses API

[njbrake]

Hi! Is there anything in progress to support the OpenResponses protocol?

https://www.openresponses.org/

It's the OpenAI sponsored open source protocol, already adopted by Huggingface as well. It's intended to replace the chat completions endpoint and become a new standard for agentic coordination with an LLM.

There's a similar issue created in llama.cpp so I thought it might be good to have it in mlx-lm as well as llama.cpp.

ggml-org/llama.cpp#19138

Thanks!