Issue with Deploying GPT-4o 2024-08-06 Model in Azure OpenAI

Question

I am using Azure OpenAI in the EUS2 location. In the model catalog, I can find the GPT-4o model dated 2024-08-06. However, when I try to deploy this model, I am unable to locate the 2024-08-06 version in the dropdown menu. Has anyone else encountered this issue?

Answer 1

This is so confusing, you have even published a blog post[1] to announce the new model on Azure, but it's not actually available?

[1] https://azure.microsoft.com/en-us/blog/announcing-a-new-openai-feature-for-developers-on-azure/

Answer 2

Mohit Garg Greetings & Welcome to Microsoft Q&A forum!

I am using Azure OpenAI in the EUS2 location. In the model catalog, I can find the GPT-4o model dated 2024-08-06. However, when I try to deploy this model, I am unable to locate the 2024-08-06 version in the dropdown menu. Has anyone else encountered this issue?

I understand that you are unable to select the version 2024-08-06 from the dropdown for deploying GPT-4o model.

I'm also not seeing the model version in drop down.

This is due to the model not yet available for deployment. When you set "Auto-update to default" it will be updated to latest version when available.

As mentioned here, Azure customers can test out GPT-4o 2024-08-06 today in the new AI Studio early access playground (preview).

Unlike the previous early access playground, the AI Studio early access playground (preview) does not require you to have a resource in a specific region.

To test out GPT-4o 2024-08-06, sign-in to the Azure AI early access playground (preview) using this link.

Once you click on the link, you will be able to access the play ground.

I hope this helps. Do let me know if you have any other queries.

If the response helped, please do click Accept Answer and Yes for was this answer helpful.

Doing so would help other community members with similar issue identify the solution. I highly appreciate your contribution to the community.

Answer 3

Thanks, Ashok, for the information, but it doesn't look like the playground is using the latest version of the model which allows a 16k output. I am getting an error indicating that the token limit is exceeded: "Max size: 4000 tokens."

Is there any way to increase max tokens settings?