Codex error
How to fix "Selected model is at capacity. Please try a different model."
Recently, many people using Codex have hit "Selected model is at capacity. Please try a different model." and cannot continue their task. Below is what people are seeing, why it happens, and fixes to try in order.
What people are seeing
The error tends to hit many people in the same hours; some cannot use their official account while a third-party relay still works. Switching IP helped a few people, but users in many countries hit it too, so the network is not the main cause; it looks more like a capacity problem on OpenAI's side.
It is not always a full stop: sometimes Codex answers a line or two and then cuts off again, which looks like queueing. Some people also keep getting it after using up Pro 20x and switching to Plus, and suspect their account is being throttled; that is only a guess so far.
Cause
OpenAI does not have enough capacity for the model, and any account can hit this. If it keeps happening only on your account for a long time, the account may also be limited.
Solutions
These go from lightest to heaviest. Try them in order.
- Retry a few times
Click continue or resend. It sometimes gets through, but the answer quality is not guaranteed.
- Remove sessions you do not recognise
On the web, open Settings → Security and log out any active session you do not recognise. One user found a session they did not recognise here and recovered after logging it out.
- Log out of all sessions and sign in again
Log out everywhere, then sign back in. Some users recovered this way.
- Change password, re-enable two-factor, wait about 3 hours
Change your password, turn two-factor authentication off and on again, log out of all sessions, then wait about 3 hours before signing in again.
- Appeal to OpenAI
If nothing else works, appeal to OpenAI: say your usage does not violate policy and ask to have normal access restored (links and a sample message below). Some users stopped seeing the error after a successful appeal; others saw it come back later.
Appeal message
A message one user sent before their account recovered. Use it as a starting point, and add the exact error text and that your usage does not violate policy.
I've been using GPT for a long time, and only recently started noticing what feels like a real drop in quality or intelligence. I've always trusted you guys to provide the level of service and quality users should be able to expect. I also genuinely believe that better collaboration between people and AI can help us do better work and, in some small way, make the world a better place.
So, if there's any chance you can press the 'back to normal effort/model' button on your side, I'd be incredibly grateful.Tips
- The first three steps take a minute and change nothing on your account.
- If it comes back after recovering, repeat from step 1 before appealing again.