Timeout errors for OpenAI gpt-4o (08-06) in region US East

Question

Timeout errors for OpenAI gpt-4o (08-06) in region US East

schoell 50

I am experiencing timeout errors for my deployment of gpt-4o (Version 2024-08-06) in region east us with API version 2024-10-21.

All my requests are answered with a timeout error. This behavior started at 07:15 AM (UTC).

Shaik Mohammad Abdul Mateen 0 Reputation points

2025-02-24T16:16:18.88+00:00

same here, tried accessing it differently, but still had no luck.
Simon-Pierre Tessier 30 Reputation points

2025-02-24T18:19:50.91+00:00

We are having the same issue with a regional US-EAST GPT4o deployment. Microsoft hasn't acknowledge the issue anywhere which is frustrating.
Anthony Helou 5 Reputation points

2025-02-24T19:46:48.6933333+00:00

We are having the same issues with version 2024-08-06 that was affecting production resources. Ended up creating a new 4o deployment with the 2024-11-20 version and everything seems to be fine now.
Dillinger 10 Reputation points

2025-02-24T19:49:08.9966667+00:00
This is affecting my organization as well. Same issues. 408 errors.

Request exceeded the timeout sent in the request: 10000ms

We even have an automatic fallback to hit gpt-4 0613, those eventually hit the same timeouts.

It would really be great if the Service Health dashboards meant anything to Microsoft.
Saideep Anchuri 9,500 Reputation points Moderator

2025-02-26T08:21:35.28+00:00

HI schoell

Following up to see if the above answer was helpful.

Thank You.
Saideep Anchuri 9,500 Reputation points Moderator

2025-02-27T08:02:20.29+00:00

HI schoell

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet

Thank You.

4 answers

Your answer

Shaik Mohammad Abdul Mateen 0 Reputation points

2025-02-24T16:16:18.88+00:00

same here, tried accessing it differently, but still had no luck.
Simon-Pierre Tessier 30 Reputation points

2025-02-24T18:19:50.91+00:00

We are having the same issue with a regional US-EAST GPT4o deployment. Microsoft hasn't acknowledge the issue anywhere which is frustrating.
Anthony Helou 5 Reputation points

2025-02-24T19:46:48.6933333+00:00

We are having the same issues with version 2024-08-06 that was affecting production resources. Ended up creating a new 4o deployment with the 2024-11-20 version and everything seems to be fine now.
Dillinger 10 Reputation points

2025-02-24T19:49:08.9966667+00:00

This is affecting my organization as well. Same issues. 408 errors.

Request exceeded the timeout sent in the request: 10000ms

We even have an automatic fallback to hit gpt-4 0613, those eventually hit the same timeouts.

It would really be great if the Service Health dashboards meant anything to Microsoft.
Saideep Anchuri 9,500 Reputation points Moderator

2025-02-26T08:21:35.28+00:00

HI schoell

Following up to see if the above answer was helpful.

Thank You.
Saideep Anchuri 9,500 Reputation points Moderator

2025-02-27T08:02:20.29+00:00

HI schoell

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet

Thank You.

Answer 1

Ahzam Fawmee 10

I encountered the same issue with gpt-4o (Version 2024-08-06) in the East US region. This seems to be a recurring pattern where older model versions experience degraded performance or timeouts once a new version is released.

In my case, updating to the latest available model version (2024-11-20) resolved the issue immediately. The latest version is significantly faster and more responsive. If you haven’t already, I’d recommend switching to it to resolve the timeout errors.

Dillinger 10 Reputation points

2025-02-25T14:30:15.43+00:00
Thanks.
My issue isn't with your resolution (this makes sense), more with Microsoft's approach to shadow-deprecating model versions:

2024-08-06 is barely 6 months ago.

My deployments use the Once a new default version is available version update policy, but 2024-11-20 is not the default. So there is no automation setting that ensures we remain on a model version that Microsoft is actually no longer supporting (e.g. 2024-08-06).

Sounds like we're all basically stuck in a never-ending (and boring) game of catch-up, making tiny 8-character changes to a model config.

Anyone have any tips to get alerted to drops of new GA model versions?
Dillinger 10 Reputation points

2025-02-25T14:30:51.74+00:00

(duplicate)

Answer 2

Baraa Ali-ElSafty 0

The same here and affected our production services

Domen Zukovec 0 Reputation points

2025-02-24T09:18:42.57+00:00

Same here, had to change the model version to fix it...
Baraa Ali-ElSafty 0 Reputation points

2025-02-24T09:26:21.9233333+00:00

What is the version that is working now?
Domen Zukovec 0 Reputation points

2025-02-25T09:20:56.2033333+00:00

(version:2024-05-13) works for me

Answer 3

Tiến Đạt Dương 20

We have same issue but Azure has yet to officially respond to the incident.

Rohit Shetty 5 Reputation points

2025-02-25T04:32:18.24+00:00

we are having the same issue, do you all think if we switch o1 mini can be a good back up strategy? Curious to hear what methods others have applied for production - is there a back up plan than can be implemented?

Answer 4

HI schoell

I understand that you are encountering an issue with your GPT-4o deployment.

Here are some steps:

1.Sometimes, clearing your browser cache can help resolve timeout issues.

Consider deploying your model in an alternative region.
Visit the Service Health - Microsoft Azurepage to see if there are any known issues in the US East region.
Communication between my system and your browser may momentarily disconnect, causing drops.

5.Add on changing Api version to previous GA version and other model version.

Kindly refer below Link: API version lifecycle

Reference Thread: : OPEN AI (GPT-4o-mini) in EAST US not working - Microsoft Q&A

Thank You.

Share via

Timeout errors for OpenAI gpt-4o (08-06) in region US East

4 answers

Your answer