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

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

• #28303
• #27149

Powered by Codex Action

Same here: 019ecf29-4e66-7560-9d34-119832dbecea
ChatGPT 5.5 extra high. ChatGPT Pro 20x plan. @openai/codex cli client on Dev Containers (Fedora Sliverblue 44).

me too： ChatGPT 5.5 extra high. ChatGPT Pro 20x plan.

Oh. This is: https://status.openai.com/incidents/01KV7ZT644J4V94GSXMFPY2ANR

me too： ChatGPT 5.5 extra high Selected model is at capacity. Please try a different model.. me ChatGPT Pro 20x plan.

me too： ChatGPT 5.5 extra high Selected model is at capacity. Please try a different model.. me ChatGPT Pro 5x plan.

Same issue here on Codex app / macOS.

Error text: Selected model is at capacity. Please try a different model.

Observed time: 2026-06-16 20:11 CST / 12:11 UTC.

Environment:

• Codex app bundled CLI: codex-cli 0.140.0-alpha.2
• Standalone CLI on PATH: codex-cli 0.130.0
• Platform: macOS

Impact: recurring interruptions during normal Codex app usage; switching/retrying still hits the capacity message. This appears to match the active OpenAI status incident: https://status.openai.com/incidents/01KV7ZT644J4V94GSXMFPY2ANR

I can submit in-product /feedback separately if needed so logs/session ID are attached.

Adding another data point with explicit quota/account details.

Environment:

• Codex Desktop App: 26.609.71450
• macOS: 26.5.1
• Subscription: ChatGPT Pro 20x
• Observed date: 2026-06-16

Models affected:

• GPT-5.5 High
• GPT-5.5 Extra High / XHigh

Error shown repeatedly:

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

Important detail: my account still shows plenty of remaining usage in both the 5-hour allowance and the weekly allowance. This does not look like normal quota exhaustion from the user side, but the error is shown repeatedly anyway. It affects both GPT-5.5 High and GPT-5.5 XHigh.

Expected behavior:

• If this is regional/model-serving capacity rather than account quota exhaustion, the UI/error message should make that distinction clear.
• If a Pro 20x user still has substantial 5-hour and weekly allowance remaining, Codex should either serve the selected GPT-5.5 model, retry with backoff, or provide a clearer availability/region/capacity explanation.
• The current message makes it look like switching models is the only remedy, but both 5.5 High and 5.5 XHigh can fail with the same error.

I do not have an in-product /feedback ID from this reporting path, but this is reproducible repeatedly today on the account described above.

As already mentioned they're working on it

https://status.openai.com/incidents/01KV7ZT644J4V94GSXMFPY2ANR

font{
line-height: 1.6;
}
ul,ol{
padding-left: 20px;
list-style-position: inside;
}

Ok, thx!

---- Replied Message ----

From

Can @.***>

Date

06/16/2026 20:32

To

@.***>

Cc

@.***>
,

@.***>

Subject

Re: [openai/codex] Selected model is at capacity. Please try a different model. (Issue #28507)

CanRau left a comment (openai/codex#28507)
As already mentioned they're working on it
https://status.openai.com/incidents/01KV7ZT644J4V94GSXMFPY2ANR

—Reply to this email directly, view it on GitHub, or unsubscribe.Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS and Android. Download it today!
You are receiving this because you commented.Message ID: @.***>

Additional data point from the same macOS Codex app environment:

The interruption is now showing as a usage-limit block even though the in-app usage panel appears to show remaining allowance.

Visible in the screenshot:

• Usage panel title: 剩余用量 / remaining usage
• 5小时: 75%, reset/time shown as 00:57
• 1周: 96%, reset/date shown as 6月23日
• Conversation error banner: You've hit your usage limit. Add credits to continue, or try again at 2026年6月17日 00:57.

This makes the failure look inconsistent: the task is interrupted as if the account hit a usage limit, while the usage UI still indicates available quota / remaining usage. This may be related to the active capacity incident, but the user-facing error now points to account limits rather than model capacity.

same for me throughout the whole day
019ecad6-cb71-70b1-a2c1-d42c05fe915f

still having this problem, despite being marked as resolved on status webpage

I’m seeing the same issue on Codex Desktop.

Environment:

• OS: Windows
• Codex Desktop app version: 26.623.61825
• Subscription: ChatGPT Plus
• Model: GPT-5.5
• Time of occurrence: around 6:00 PM KST
• Remaining usage:
• 5-hour limit: 87% remaining
• 1-week limit: 81% remaining

Error banner:

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

Conversation error:

stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID 510f4cac-b03c-403f-8598-8a2064c5a115 in your message.

Request ID:

510f4cac-b03c-403f-8598-8a2064c5a115

This happened despite having remaining usage quota, so it was unclear whether the failure was due to temporary GPT-5.5 capacity, quota handling, or a Codex Desktop streaming issue.

Same for me! (occured right after update)

OS: MacOS
Codex Desktop app version: 26.623.61825
Subscription: ChatGPT Plus
Model: GPT-5.5
Time of occurrence: around 12:50 CEST
Remaining usage:
5-hour limit: 99% remaining
1-week limit: 100% remaining

Request ID: fc05a3a6-29f6-4c12-9467-a9bf8fb7339e

stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID fc05a3a6-29f6-4c12-9467-a9bf8fb7339e in your message.

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

OS: MacOS
Codex Desktop app version: 26.623.61825

Subscription: ChatGPT Plus
Model: GPT-5.5 High
Request ID d10095e3-5ca7-4310-a56a-377626c7a1f2

stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID d10095e3-5ca7-4310-a56a-377626c7a1f2 in your message.

Is this a global issue today? I am also getting the same error "Selected model is at capacity. Please try a different model." with no clear indication whether it is a server overload, or tokens exhaustion, yet I am at ~200/353 tokens used. I am on Pro Plan.

I have the same issue, Selected model is at capacity. Please try a different model.
OS: Windows 11
Codex Desktop app version: 26.623.61825
Subscription: ChatGPT Plus
Model: GPT-5.5 High

Yesterday, before upgrading codex to latest version, it was normal. However, when upgrading to latest version, the issue came up.

Same issue here.

OS: macOS (M1)
Codex Desktop version: 26.623.61825
Subscription: ChatGPT Plus
Model: GPT-5.5 Medium

Error:

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

Followed by:

"stream disconnected before completion: An error occurred while processing your request."

This is also occurring in the Codex CLI, so it does not appear to be limited to the desktop application.

Remaining usage quota is available, so this does not appear to be a usage-limit issue.

After codex desktop updata.

OS: macOS (M4)
Codex Desktop version: 26.623.61825
Subscription: ChatGPT Plus
Model: GPT-5.5

Error:

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

Followed by:

stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID 35989ca7-4242-4e49-8ba4-a411919e325e in your message.

+1

+1

+1

Selected model is at capacity. Please try a different model. +1

Hi OpenAI Support,
I’m repeatedly seeing this error while using Codex Desktop on a Pro account:

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

This happened in the Codex thread named:
Review WhatsApp service
Relevant identifiers:

Conversation/session ID:
019dc994-0193-7080-83dd-b2ed882ab927

Desktop app session IDs seen in local logs:
92070f11-cd11-48ad-ab63-38315782e48e
ce887626-77c8-4a6a-8859-f1b9ef950f8a

Relevant timestamps:
2026-07-07T08:23:18Z
Latest turn ID: ff91704d-99c1-415b-b374-dced14b1e247
Status: interrupted

2026-07-08T09:12Z
Turn ID: 3f7c6186-bf38-49bc-ba76-047cf8f61bda

My local logs show plan_type: "pro" and rate_limit_reached_type: null, so this does not appear to be a normal account rate-limit issue from the local log data.

Could you please investigate whether these capacity errors were caused by backend model availability or routing for this Codex session?

Thank you.

Same issue.

M5 Max, Plus, 5.5 medium

Does not seem to be due to rate limits.

Session ID: 019f3987-3b79-70e0-85a6-41c07c1bb2aa

Please investigate this issue. Thanks!

same here 5.6 sol on high effort

Same issue happened on 5.6 Sol Ultra
Pleae look into it, thanks

happening a lot now

BTW it's a great way of performing a real world load testing 😁

Happened to me with 5.6 Sol on X-high in Goal Mode, and I'd never seen it before. PRO 20x user with plenty of usage allotment for daily and weekly with multiple resets.

Same Happened to me with 5.6 Sol on X-high