Codex CLI use gpt-5.1-codex-mini
Resolved 💬 7 comments Opened Feb 12, 2026 by notyesbut Closed Feb 15, 2026
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex CLI is running?
Codex CLI: 0.99.0-alpha.10
What subscription do you have?
ChatGPT Pro (also completed OpenAI KYC / Trusted Access for Cyber)
Which model were you using?
gpt-5.3-codex xhigh
What platform is your computer?
Linux 6.6.87.2-microsoft-standard-WSL2 x86_64 x86_64
What terminal emulator and version are you using (if applicable)?
VSCode/WSL
What issue are you seeing?
Evidence from codex-tui.log:
TUI logs show model selection:
- 2026-02-06T22:19:50Z INFO ... Selected model: gpt-5.3-codex, Selected effort: xhigh
- 2026-02-06T23:18:11Z INFO ... Selected model: gpt-5.3-codex, Selected effort: xhigh
- 2026-02-09T01:49:38Z INFO ... Selected model: gpt-5.3-codex, Selected effort: xhigh
But SSE response.model never uses gpt-5.3-codex.
Full log scan of response.model:
- 43x gpt-5.1-codex-mini
- 6x gpt-5.2-2025-12-11
- 0x gpt-5.3-codex
What steps can reproduce the bug?
Command:
rg --text -o '"model":"[^"]+"' ~/.codex/log/codex-tui.log | sort | uniq -c | sort -nr
What is the expected behavior?
_No response_
Additional information
_No response_
7 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Are you using an Azure endpoint?
Please use the
/feedbackslash command to upload your logs, and post the thread ID here so I can take a look.Closing because we don't have enough details to diagnose. @notyesbut, if you can provide the above requested information, I can take a look.
thread ID 019c6346-b76a-7f90-b304-834892b49bfe in an existing issue.
https://github.com/openai/codex/issues/new?template=2-bug-report.yml&steps=Uploaded%20thread:%20019c6346-b76a-7f90-b304-834892b49bfe
Thanks for providing the logs. I don't see any evidence of gpt-5.1-codex-mini being used. All SSE responses in the log are reporting responses from "gpt-5.3-codex", which is what was selected.
If you are seeing responses from gpt-5.2-2025-12-11, that is explained here.
Ok, thanks, it seems this has already been fixed.