Cannot deploy GPT model in southeastasia region

NNG Admin 0

User's image

Hi everyone,
I am trying to deploy the GPT-4 model in the Southeast Asia region. According to the documentation provided by Microsoft, GPT-4 can be deployed in Southeast Asia. However, when I choose 'Global Provisioned-managed,' there is no available quota for deployment. Could you please confirm whether it is possible to deploy an LLM in Southeast Asia? Has anyone previously deployed a GPT model in Southeast Asia?
Your help would be greatly appreciated.
Thank you very much!

SriLakshmi C 3,015 Reputation points Microsoft External Staff

2025-03-05T09:49:07.0166667+00:00

Hello NNG Admin,

Greetings and Welcome to Microsoft Q&A!

The GPT-4 model is not supported for deployment in the Southeast Asia region, as confirmed by Microsoft's official documentation Global standard model availability. However, the GPT-4o model is supported in this region and can be successfully deployed. If you are looking to deploy a Large Language Model (LLM) in Southeast Asia, you should consider using GPT-4o instead of GPT-4.

Based on testing and verification, GPT-4o can be deployed using the Global Provisioned-managed option, which allows for a more flexible allocation of resources across multiple regions. If you encountered quota limitations or unavailability when trying to deploy GPT-4, switching to GPT-4o should resolve the issue.

If you specifically require GPT-4, you will need to select a different region where it is supported, such as East US or West Europe. However, for Southeast Asia, GPT-4o is the recommended alternative.

I hope you understand. And, if you have any further query do let us know.

Thank you!
NNG Admin 0 Reputation points

2025-03-05T12:25:18.71+00:00

Hello SriLakshmi C,

Thank you very much, I sincerely appreciate your information.
This is my mistake - I intended to deploy GPT-4o, not GPT-4, in Southeast Asia. However, as you can see in the screenshot of the question part, I am trying to deploy GPT-4o with the Global Provisioned-Managed option, but it still shows 'No quota.' Do I need to request additional quota?
I also have questions:
Do I have to pay for a quota request, or do I only pay for the quota that I actually use?
Does the Global Provisioned-Managed option provide a 'Pay as you go' option?
Based on the document, it only offers the option to charge by the hour(image of document below).
Thank you for reading. I look forward to hearing from you.
NNG Admin 0 Reputation points

2025-03-05T12:32:12.53+00:00

Hi SriLakshmi C,
Thanks for your response. I sincerely appreciate your information.
This is my mistake—I intended to deploy GPT-4o, not GPT-4, in Southeast Asia. However, as you can see in the screenshot of the question part, I am trying to deploy GPT-4o with the Global Provisioned-Managed option, but it still shows 'No quota.' Do I need to request additional quota?
I also have questions:
Do I have to pay for a quota request, or do I only pay for the quota that I actually use? And does the Global Provisioned-Managed option provide a 'Pay as you go' option?
Based on the document, it only offers the option to charge by the hour.
NNG Admin 0 Reputation points

2025-03-06T06:52:53.64+00:00

Hi SriLakshmi C,
Could you help me with this problem?
Thanks you very much
NNG Admin 0 Reputation points

2025-03-06T07:22:40.2333333+00:00

Hi SriLakshmi C,
Could you help me with this problem please?
SriLakshmi C 3,015 Reputation points Microsoft External Staff

2025-03-06T09:29:07.8133333+00:00

Hi NNG Admin,

I have also encountered this issue at times. Please try deleting the deployment, then recreate and deploy it again. If the issue persists, wait for some time and try again. Additionally, refreshing the page may help resolve the problem.

I was able to reproduce the issue in my environment, and the GPT-4o model is successfully deploying under Global Provisioned-Managed. Here is the screenshot for reference.

Do I have to pay for a quota request, or do I only pay for the quota that I actually use?

You don’t have to pay for submitting a quota request, you only pay for the quota you actually use.

And does the Global Provisioned-Managed option provide a 'Pay as you go' option?

The Global Provisioned-Managed option does not follow a traditional "Pay-as-you-go" model. Instead, it operates on an hourly billing system, where charges are based on the number of Provisioned Throughput Units (PTUs) deployed. This means that once a model is deployed, costs are calculated based on the hourly rate multiplied by the number of PTUs in use.

please refer this Hourly usage, Azure OpenAI provisioned Managed offering updates.

Thank you!
SriLakshmi C 3,015 Reputation points Microsoft External Staff

2025-03-07T09:06:41.08+00:00

Hi NNG Admin,

Did you get any chance to check the above response. Thank you!
NNG Admin 0 Reputation points

2025-03-07T15:56:16.87+00:00

Hi SriLakshmi C,
I am still encountering the 'No quota' . Could you show me your quota? My quota usage/limit is currently 0, and I believe this is the reason I am unable to deploy the model. Could you help me clarify this?
Thank you very much for your help."

Share via

Cannot deploy GPT model in southeastasia region

Your answer