[FEATURE] Add support for `gpt-5.2-codex` model

[compumike]

Scope check

• This is core LLM communication (not application logic)
• This benefits most users (not just my use case)
• This can't be solved in application code with current RubyLLM
• I read the Contributing Guide

Due diligence

• I searched existing issues
• I checked the documentation

What problem does this solve?

The OpenAI model gpt-5.2-codex is listed in the model registry. These codex models are trained for programming-specific performance and can be useful to chat with for programming-related use cases.

However, when I try to use it with RubyLLM, I get an error:

chat = RubyLLM.chat(model: "gpt-5.2-codex")
=>                           
#<RubyLLM::Chat:0x00007b6467b7c4f0
...                          

chat.ask("Write a Python function that adds two numbers.")
in '<main>': This is not a chat model and thus not supported in the v1/chat/completions endpoint. Did you mean to use v1/completions? (RubyLLM::Error)

(The above works fine if I use model: "gpt-5.2". The problem is specifically with model: "gpt-5.2-codex".)

Proposed solution

I'm not sure what the easiest implementation in RubyLLM would be.

However, I do know that it is possible to have a multiturn chat experience with gpt-5.2-codex when using the official openai gem. But doing so gives up all the convenient developer affordances that RubyLLM provides, so I'd rather be able to use this model within RubyLLM! 😄

Why this belongs in RubyLLM

It's still multiturn text conversations which fit the RubyLLM.chat concept. I have a feeling that codex type use cases are desired by other RubyLLM users.

[compumike]

@tpaulshippy Did some more research... is this feature request related to (or would be solved solved by) #213 ?

[tpaulshippy]

@tpaulshippy Did some more research... is this feature request related to (or would be solved solved by) #213 ?

Yes. As I mentioned in #325 certain models are only available via the responses API. If you want to make this work now with RubyLLM some options are https://github.com/khasinski/ruby_llm-responses_api and https://github.com/tpaulshippy/ruby_llm_community

[compumike]

@tpaulshippy Perfect, thank you! ❤️

[crmne]

Closing as a duplicate of #213.

Support for gpt-5.2-codex depends on OpenAI Responses API support tracked in #213.