Azure AI Foundry Base Model Deployments

Question

Azure AI Foundry Base Model Deployments

Nico Lim 0

I keep having this error whenever I try to deploy Mistral models in Azure AI Foundry. Initially my deployments of gpt-4o and grok-3 have this issue but I just let it rest for a few days and trying to deploy again and it worked, but this mistral model never gets to be deployed. Is Mistral region locked or something? This happens to all mistral models. User's image

Ravada Shivaprasad 1,115 Reputation points Microsoft External Staff Moderator

2025-08-07T19:38:08.85+00:00

Hi Nico Lim

The error you're encountering when deploying Mistral models in Azure AI Foundry—specifically the message stating “Marketplace Subscription purchase eligibility check failed.... The Publisher does not make available Offer 'mistral-large-2411'... in your Subscription/Azure account's region”—indicates that the Mistral models are currently region-restricted. Unlike GPT-4o and Grok-3, which may have broader regional availability.

Mistral models such as mistral-large-2411 are not offered in all Azure regions, including East US, where your deployment is configured. This restriction is enforced by the Azure Marketplace and depends on the publisher’s configuration for each model. To resolve this, you can either switch your deployment to a supported region (e.g., West Europe or Central US) or contact Azure support to request access in your current region.
The inconsistent availability pattern you've observed is documented behavior, where model availability can change daily, and some users have reported being able to deploy models in certain regions on some days but not others.

While waiting periods sometimes resolve the issue (as seen with gpt-4o and grok-3), Mistral models specifically tend to show more persistent regional restrictions. If you continue experiencing issues after trying these steps, reaching out to Azure support would be the most direct path to resolution.

Reference links : mistral-ocr-now-available , Region availability

Hope it helps!

Thank you
Nico Lim 0 Reputation points

2025-08-11T02:46:18.6266667+00:00

I see, I understand. Although from the reference link you provided on region availability. It is stated there that mistral models are available on East US hub/project region for deployment. is it some kind of mistake in the documentation?
Manas Mohanty 9,655 Reputation points Microsoft External Staff Moderator

2025-08-13T09:21:05.6033333+00:00

Hi Nico Lim

These errors can also be rising due to regional capacity constraints at that time.

Best option will be try in other supported regions.

Assuming that it is a paid subscription with valid payment method as Sponsored subscription or Trial subscription won't be eligible for deploying models from model catalogue.

Thank you.
Manas Mohanty 9,655 Reputation points Microsoft External Staff Moderator

2025-08-15T09:41:33.15+00:00

Hi Nico Lim

Please let us know switching to other supported regions helped address the issue.

Thank you
Nico Lim 0 Reputation points

2025-08-18T12:20:42.2733333+00:00

Hi Manas,

Sorry for not getting back at the earliest. I'm not sure if we are allowed to switch to other regions after deployment of the resource group. the option to move the azure AI foundry capacity is greyed out.

Regards,
Nico
Ravada Shivaprasad 1,115 Reputation points Microsoft External Staff Moderator

2025-08-28T17:36:53.9433333+00:00

Hi Nico Lim

Sorry for Delayed response

Once a resource group is deployed in Azure AI Foundry, its capacity becomes region-bound, meaning it cannot be moved to another region. This is a known limitation, and the option to switch regions is intentionally disabled post-deployment. If you’re seeing the region-switching option greyed out, it’s because Azure does not support relocating AI Foundry capacity across regions once the resource group is active.

To work in a different region, you’ll need to create a new resource group in that region and manually reconfigure your AI Foundry setup, including migrating models and configurations. For temporary testing in other regions, Azure offers a shared quota pool that allows limited access to capacity across regions, but this is not suitable for production workloads.

Hope it helps!

Thank you

Your answer

Ravada Shivaprasad 1,115 Reputation points Microsoft External Staff Moderator

2025-08-07T19:38:08.85+00:00

Hi Nico Lim

The error you're encountering when deploying Mistral models in Azure AI Foundry—specifically the message stating “Marketplace Subscription purchase eligibility check failed.... The Publisher does not make available Offer 'mistral-large-2411'... in your Subscription/Azure account's region”—indicates that the Mistral models are currently region-restricted. Unlike GPT-4o and Grok-3, which may have broader regional availability.

Mistral models such as mistral-large-2411 are not offered in all Azure regions, including East US, where your deployment is configured. This restriction is enforced by the Azure Marketplace and depends on the publisher’s configuration for each model. To resolve this, you can either switch your deployment to a supported region (e.g., West Europe or Central US) or contact Azure support to request access in your current region.
The inconsistent availability pattern you've observed is documented behavior, where model availability can change daily, and some users have reported being able to deploy models in certain regions on some days but not others.

While waiting periods sometimes resolve the issue (as seen with gpt-4o and grok-3), Mistral models specifically tend to show more persistent regional restrictions. If you continue experiencing issues after trying these steps, reaching out to Azure support would be the most direct path to resolution.

Reference links : mistral-ocr-now-available , Region availability

Hope it helps!

Thank you
Nico Lim 0 Reputation points

2025-08-11T02:46:18.6266667+00:00

I see, I understand. Although from the reference link you provided on region availability. It is stated there that mistral models are available on East US hub/project region for deployment. is it some kind of mistake in the documentation?
Manas Mohanty 9,655 Reputation points Microsoft External Staff Moderator

2025-08-13T09:21:05.6033333+00:00

Hi Nico Lim

These errors can also be rising due to regional capacity constraints at that time.

Best option will be try in other supported regions.

Assuming that it is a paid subscription with valid payment method as Sponsored subscription or Trial subscription won't be eligible for deploying models from model catalogue.

Thank you.
Manas Mohanty 9,655 Reputation points Microsoft External Staff Moderator

2025-08-15T09:41:33.15+00:00

Hi Nico Lim

Please let us know switching to other supported regions helped address the issue.

Thank you
Nico Lim 0 Reputation points

2025-08-18T12:20:42.2733333+00:00

Hi Manas,

Sorry for not getting back at the earliest. I'm not sure if we are allowed to switch to other regions after deployment of the resource group. the option to move the azure AI foundry capacity is greyed out.

Regards,
Nico
Ravada Shivaprasad 1,115 Reputation points Microsoft External Staff Moderator

2025-08-28T17:36:53.9433333+00:00

Hi Nico Lim

Sorry for Delayed response

Once a resource group is deployed in Azure AI Foundry, its capacity becomes region-bound, meaning it cannot be moved to another region. This is a known limitation, and the option to switch regions is intentionally disabled post-deployment. If you’re seeing the region-switching option greyed out, it’s because Azure does not support relocating AI Foundry capacity across regions once the resource group is active.

To work in a different region, you’ll need to create a new resource group in that region and manually reconfigure your AI Foundry setup, including migrating models and configurations. For temporary testing in other regions, Azure offers a shared quota pool that allows limited access to capacity across regions, but this is not suitable for production workloads.

Hope it helps!

Thank you

Share via

Azure AI Foundry Base Model Deployments

Your answer