Separate model selection for Plan vs Execute modes (e.g., GPT-5-high for planning + GPT-5-Codex execute)

### What feature would you like to see?

Allow Codex to use one model for planning and a different one for execution. Example: use a gpt5-high to outline a solution, then switch to a gpt5-codex to implement.

Note : i know plan mode does not exist yet. 

### Additional information

_No response_