Codex App used with Azure hosted models

What version of the Codex App are you using (From “About Codex” dialog)?

26.226.940 (720)

What subscription do you have?

Azure models proxied via LiteLLM

What platform is your computer?

Darwin 24.6.0 arm64 arm

What issue are you seeing?

I've set up the App to use our Azure models proxied via LiteLLM, works via

model = "gpt-5.3-codex"
model_provider = "litellm"
model_reasoning_effort = "high"

[model_providers.litellm]
name = "OpenAI"
base_url = "<redacted>"
wire_api = "responses"
env_key = "OPENAI_API_KEY"

in my ~/.codex/config.toml.

Two issues I see:

• no model selection dropdown, no effort selection
• CLI allows me switching between the models, if LiteLLM doesn't allow particular model for my key, usage of it simply errors out
• agent does not resume work between turns, eg it does some exploration, writes the intermediate progress message and stops instead of continuing (the message clearly states the agent should continue).
• prompting it to retry resumes work

What steps can reproduce the bug?

• set up the App towards Azure/LiteLLM proxied models
• no model or effort selection in the app
• agent stops between turns

What is the expected behavior?

• model and effort selection is available
• agent can work without interruptions

Additional information

_No response_