Selecting `gpt-5.3-codex` in Codex shows response.model = `gpt-5.2-2025-12-11`

[UniverseFly]

What version of Codex CLI is running?

codex-cli 0.98.0

What subscription do you have?

Pro

Which model were you using?

gpt-5.3-codex

What platform is your computer?

Darwin 24.1.0 arm64 arm

What terminal emulator and version are you using (if applicable)?

Terminal.app

What issue are you seeing?

When I select gpt-5.3-codex in Codex, the returned Responses API object reports a different model: gpt-5.2-2025-12-11.

Log evidence (PII redacted) from ~/.codex/log/codex-tui.log:

2026-02-05T21:02:42.597590Z  INFO  codex_tui::chatwidget: Selected model: gpt-5.3-codex, Selected effort: xhigh

2026-02-06T21:34:23.110590Z TRACE codex_api::sse::responses: SSE event: {"type":"response.completed","response":{"model":"gpt-5.2-2025-12-11", ...}}

What steps can reproduce the bug?

1. Run Codex with SSE tracing enabled:

   RUST_LOG='codex_tui::chatwidget=info,codex_api::sse::responses=trace' codex

2. In the model picker, select gpt-5.3-codex (effort xhigh).
3. Send any prompt (e.g. “hello”).
4. Inspect ~/.codex/log/codex-tui.log and find the response.completed SSE event; observe response.model is gpt-5.2-2025-12-11.

What is the expected behavior?

If I choose gpt-5.3-codex, I expect response.model reflects gpt-5.3-codex.

Additional information

No response

[etraut-openai]

If you're able to repro this, please use /feedback to upload your logs and session details, then post the thread ID here.

[UniverseFly]

If you're able to repro this, please use /feedback to upload your logs and session details, then post the thread ID here.

Thanks for the guidance. Here is the thread id: 019c3695-daf8-7b11-b5df-f53832fb233d.

[etraut-openai]

Have you continued to see this problem, or was this a one-time issue?

[UniverseFly]

Yes, the issue still existed in my try just now. Is there a way to verify which model is routed for inference in the backend? My guess is either incorrect routing (the underlying model isn't gpt-5.3-codex) or a less severe missing mapping that triggers a fallback model name in the final response (the underlying model is still gpt-5.3-codex).