Codex Desktop on Windows rejects GPT-5.5 with "This model is not supported when using X-OpenAI-Internal-Codex-Responses-Lite."

Resolved 💬 1 comment Opened Jun 26, 2026 by MrBear999-ui Closed Jun 26, 2026

What version of Codex?

Codex Desktop / CLI version 0.142.2

Operating system

Windows

What happened?

Selecting gpt-5.5 in Codex Desktop consistently fails before any assistant response is produced.

The UI shows this error:

This model is not supported when using X-OpenAI-Internal-Codex-Responses-Lite.

I reproduced this after fully restarting the app and opening a new window.

Steps to reproduce

  1. Open Codex Desktop on Windows.
  2. Start or continue a thread.
  3. Switch the model to gpt-5.5.
  4. Send a prompt.
  5. Observe that the request fails immediately and no assistant message is produced.

Expected behavior

gpt-5.5 should run normally in Codex Desktop, or the UI should prevent selecting it if the current backend route does not support it.

Actual behavior

The app allows selecting gpt-5.5, but the request fails with the message:

This model is not supported when using X-OpenAI-Internal-Codex-Responses-Lite.

Additional evidence

  • The same machine can still run gpt-5.4 successfully in the same workspace.
  • Earlier on the same date, gpt-5.5 had worked in other threads, so this does not look like a simple user misconfiguration.
  • In local session transcripts for the failing thread(s), the turn context records model: gpt-5.5, but the task completes with last_agent_message: null.

Example local evidence:

  • Thread/session id: 019f03a2-5384-7cc2-8f6a-ffaafb93a993
  • Failing session file: C:\Users\熊阔豪\.codex\sessions\2026\06\26\rollout-2026-06-26T19-13-17-019f03a2-5384-7cc2-8f6a-ffaafb93a993.jsonl
  • More failing gpt-5.5 attempts around 2026-06-26 19:27-19:28 local time also completed with no assistant output.
  • Nearby gpt-5.4 attempts in the same workspace succeeded immediately.

Impact

I currently cannot use gpt-5.5 in Codex Desktop on this Windows setup and have to fall back to gpt-5.4.

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗