How do we upgrade the retiring model gpt-35-turbo to gpt-4o-mini using a Standard deployment for regions that don't support gpt-4o-mini?

Erin Ripple 20 Reputation points Microsoft Employee
2025-03-04T20:16:01.4866667+00:00

For data boundary requirements, we are using Standard deployments for the azure open AI models. We are also open to using the data zone deployments. However, neither of these supports gpt-4o-mini in japaneast or even Asia/Pacific at large. When the gpt-35-turbo 0125 version retires May 31, 2025, will the gpt-4o-mini be available in those regions by then? According to documentation, gpt-4o-mini is the suggested replacement. Will auto upgrade automatically upgrade the model family, or does it only upgrade versions within a model family? What are the options here given this would be a serious service disruption?

Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
3,777 questions
0 comments No comments
{count} votes

Accepted answer
  1. Prashanth Veeragoni 1,190 Reputation points Microsoft External Staff
    2025-03-05T02:09:39.9733333+00:00

    Hi Erin Ripple,

    Welcome to Microsoft Q&A forum. Thanks for posting your query.As currently, gpt-4o-mini is not supported in Japan East or the Asia-Pacific region under Standard or Data Zone deployments. Since gpt-35-turbo-0125 is set to retire on May 31, 2025, Microsoft has recommended gpt-4o-mini as a replacement. However, there is no official confirmation that gpt-4o-mini will be available in those regions by the retirement date.

    Addressing Your Concerns:

    Auto Upgrade Functionality:

    The Auto Upgrade feature in Azure OpenAI automatically updates model versions within the same model family (e.g., GPT-3.5 to newer GPT-3.5 versions).

    It does not switch models across different families (e.g., GPT-3.5 to GPT-4o-mini).

    Therefore, if gpt-4o-mini is not available in your region, Auto Upgrade will not transition you to it.

    Potential Solutions to Avoid Service Disruption:

    Migrate to a supported region: Since gpt-4o-mini is currently available in select regions, you may need to consider deploying it in a region where it's supported. However, this may conflict with your data residency requirements.

    Use Global-Standard Deployment: If permitted, Global-Standard deployments allow processing across multiple Azure OpenAI regions. However, you need to verify if this aligns with your data boundary requirements.

    Wait for Microsoft Updates: Microsoft may expand gpt-4o-mini availability before May 31, 2025. It’s recommended to keep checking Azure OpenAI's official region support updates.

    You can check the latest Azure OpenAI region support updates on Microsoft's official documentation below:

    https://learn.microsoft.com/en-us/azure/cognitive-services/openai/concepts/models#model-availability

    Hope this helps. Do let us know if you any further queries.   

    ------------- 

    If this answers your query, do click Accept Answer and Yes for was this answer helpful.

    Thank you.​​

    1 person found this answer helpful.

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.