app-server supports set model_provider on `turn/start`

Resolved 💬 4 comments Opened Apr 22, 2026 by caibirdme Closed May 31, 2026

💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What variant of Codex are you using?

0.122.0

What feature would you like to see?

in thread/start, thread/resume, they both support {model, modelProvider}, while in turn/start, it just supports {model}.

Additional information

_No response_

View original on GitHub ↗

4 Comments

github-actions[bot] contributor · 4 months ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

##18764
##17541

Powered by Codex Action

etraut-openai contributor · 4 months ago

The core harness doesn't generally support switching model providers in the middle of a thread. We recommend starting a new thread if you switch providers.

Can you explain the use case that's prompting this feature request?

caibirdme · 4 months ago

The core harness doesn't generally support switching model providers in the middle of a thread. We recommend starting a new thread if you switch providers. Can you explain the use case that's prompting this feature request?

We subscribed gpt-5.4 directly from openai platform, and subscribed other gpt-x models from azure(we have discount). It's normal for people to switch model among 5.4-high,5.4-mini,5.3-codex back and forth. But the problem is, we have two providers(openai & azure)

etraut-openai contributor · 2 months ago

This feature request hasn't received enough upvotes, so I'm closing.