Custom provider Codex sessions appear affected by ChatGPT Plus usage limit
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 125k
- Forks
- 19.4k
- PR merge metrics
- PR metrics pending
Description
What version of the Codex App are you using (From “About Codex” dialog)?
0.153.4
What subscription do you have?
ChatGPT Plus
What platform is your computer?
Windows 11 x64
What issue are you seeing?
I am using Codex Desktop with a custom OpenAI-compatible provider (ToolCode), not the ChatGPT account provider.
Recently, Codex Desktop sessions using this custom provider started showing the same usage limit message and reset time as my ChatGPT Plus Work quota.
The behavior is unusual because:
- The custom provider dashboard still shows available balance and records the API requests normally.
- The provider is configured separately through an OpenAI-compatible API endpoint.
- The ChatGPT Plus 5-hour quota appears to be exhausted at the same time, with the same reset timestamp.
- Previously, this setup worked without affecting my ChatGPT Plus quota.
The Codex error message shown:
"You've hit your usage limit. Upgrade to Pro ... or try again at 1:25 PM."
The ChatGPT usage page also showed the same reset time.
What steps can reproduce the bug?
- Configure Codex Desktop with a custom OpenAI-compatible provider.
- Use the provider for normal coding tasks.
- Have a ChatGPT Plus account logged in.
- Continue using Codex until the ChatGPT Work quota is exhausted.
- Observe whether the custom provider session is also blocked by the same usage limit.
In my case, the Codex session using the custom provider displayed the same reset time as ChatGPT Plus quota.
What is the expected behavior?
Requests sent through a custom OpenAI-compatible provider should use that provider's quota and billing state.
The ChatGPT Plus subscription quota should not unexpectedly affect custom provider requests, unless this is intended behavior and documented.
Additional information
Additional observations:
- Creating a new Codex conversation sometimes allows execution again, even with the same model.
- Changing models does not reliably remove the limit.
- The issue appeared only recently after Codex configuration changes and did not occur previously.
- ToolCode usage records show the API requests were processed through the custom provider.
I am not sure whether this is:
- an intended quota sharing behavior,
- a Codex routing/account-state issue,
- or a configuration problem.
I am reporting the behavior for investigation.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the report in Codex Desktop 0.153.4 on Windows 11 with a custom OpenAI-compatible provider and a logged-in ChatGPT Plus account, then compare the Codex reset message with the provider dashboard and ChatGPT usage page. Done means determining whether custom-provider requests share ChatGPT quota, identifying the routing or account-state cause if unintended, and documenting the intended behavior if it is expected.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- api, backend
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100