Still getting a misleading message for not supported models on free accounts
Open 💬 10 comments Opened Mar 15, 2026 by claell
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What issue are you seeing?
I still get a message like {"detail":"The 'gpt-5.4' model is not supported when using Codex with a ChatGPT account."}
This is misleading and not mentioning that this only applies for free (and maybe Go?) accounts.
In #14181 and also in #14238, the bad message was acknowledged. Yet, nothing happened, so far? Is it so hard to change a string on the server side, and put a "free", or so in there? Is this planned? What is the ETA of fixing this?
What steps can reproduce the bug?
Have a ChatGPT free account in Codex and send a message to an unsupported model.
What is the expected behavior?
_No response_
Additional information
_No response_
10 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Actually, this is also not in the terminal, but in the codex app for Windows. Is this the right place to report issues or is there a different issue tracker for that in place that I could use?
I'm also seeing this suddenly today across all codex sessions, despite having a paid account!
@pirate This issue is explicitly not for paid accounts, it is for free accounts, and the error message is misleading because it implies that this would be the case for all ChatGPT accounts and not only the free accounts. So right now, the error message that is also seemingly server-side seems to be not specific, and this is what this issue is about.
@pirate, for your issue, refer to #14331.
@etraut-openai, can we get some official response to this? Is there a timeline to fix the string of the server response?
Thank you for your continued work, improving agentic coding every day!
Hi @etraut-openai, following up on this because the last team-side action I can see here is still the March 15 relabeling to
TUI, and my March 24 question has not had a response. The related account/model issues I found, such as #14412 and #18805, make this look like an account/plan-specific availability problem, but the server/API response says the model is not available for ChatGPT accounts in general. Could you clarify whether that wording is expected, or route this to whoever owns auth/model-availability error messages? Thanks for helping keep user-facing error text accurate.any update ?
@etraut-openai @aibrahim-oai @dylan-hurd-oai Pinging you here again, as it has been another normal time period and we did not get any response to this issue that should be easily solvable with a better message. I don't know whether this is still relevant even, but I hope that you are committed to providing a good user experience to all of your users. Thank you.
I am still facing it