I'm getting this error ''Selected model is at capacity. Please try a different model.''

Codex
1

I’m getting this error when calling the API:

“Selected model is at capacity. Please try a different model.”

how to handle this

2

Hi, and welcome to the developer community. I had it too like one or two times.

It means that the model is at capacity, not enough compute for so many users at that time.
I would assume it happens when a new peak level of usage is reached and they need to find a way to scale it.

Like I said it happens .. but not very often. You just have to wait or you could try to restart vs code if you use that. For me it looks like they reserve a spot for you on a model .. I think that because when i restart vs code i normally get an error message and I have to restart it again.

Many times turning it off and on again seems to work haha… I should become a 3rd level IT support guy…

3

I don’t know quite where I got this idea from, but… please try a different model when encountering an error that explains that you should handle this by trying a different model.

WIthout more GPU compute to offer: its either degrading the AI quality, or having some users try back later.

Thanks for the report that you see this - such is all over social media.

4

IMO, a better message would be:

Selected model is at capacity due to compute issues. Please try back later.

5

This error is persisting.

6

Currently experienceing this issue with… all three 5.6 models. Is the solution for this paid service going to be “use an older, worse project or wait”…? :sob:

7

You need to take some rest at some point. Whenever I experience a model not working for a few hours I grab some sleep.

8

I have this error in all models! Is it fixable or I’ll wait for answer from OpenAI ?
First time this error I see ,when working with Tiera , then Codex goes to update app, and all models are not responding

9

I’ve also noticed quite a bit of delays recently. What plan are you on?

10

I am facing quietly same issue

11

Hi!
Thank you for raising this!

Can you please use the /feedback command and share the IDs here?

12

I am in Pro plan :(((((

With early payment

13

Yes i do! Now work continues in Luna 5.6 - I’ll send it after work ends

14

019fff97-7f8e-7b12-88cb-8f9f747d996f

15

It’s happening to me today, using GPT 5.6 Sol Extra High in Coded (ChatGPT app in MacOS).

16

image
image739×150 4.46 KB

Hi,

Since August 27, 2026, Codex Desktop has repeatedly stopped active coding tasks with the following message:

Selected model is at capacity. Please try a different model.

Environment

  • Selected model: 5.6 Terra
  • Reasoning level: Medium
  • Task type: OneDrive work with reasonable access
  • Operating system: Windows
  • Approximate time: August 27, 2026, during the afternoon in Malaysia (UTC+8)

Are there anyone else from this region encountering the same problem? I see that other geographies encountered this problem before, so just wanted to confirm this is temporary and can be taken care of.

Thanks

17

Sometimes it could be some cache error or whatever if it doesn’t resolve within a reasonable period of time if it’s not too disruptive consider reinstalling the app. But first try other app solutions maybe it’s related to specific project or just a broken chat it happens rarely

18

This is happening to me repeatedly today.

19

only with terra though~~~~~~~~~~~

20

Hi!
Thank you for raising this!

Can you please use the /feedback command and share the IDs here?

This is will help the team to identify the issue when they are looking into it.