Azure OpenAI gpt-4o deployment fails with error 715-123420 on one account but works on another in the same subscription.

Deploying a gpt-4o model to an Azure OpenAI cognitive account fails with error 715-123420 on our UAT environment, while an identical configuration deploys successfully on our DEV environment. Both accounts are in the same subscription and the same region (Australia East).

Error

Performing CreateOrUpdate:

unexpected status 400 (400 Bad Request) with error: 715-123420: An error occurred.

Please reach out to support for additional assistance.

What I have tried

• Deploying via Terraform — fails
• Deploying via the Azure Portal — fails with the same error code
• Destroying and recreating the cognitive account from scratch — fails
• Deploying gpt-4o with SKU Standard instead of GlobalStandard — fails
• Deploying gpt-4o-mini — fails with the same error
• Deploying gpt-4.1-mini on the same account — succeeds

Quota

Quota for OpenAI.GlobalStandard.gpt-4o shows 125 units in use out of 450 limit, all accounted for by the DEV deployment. UAT has no active deployments.

Conclusion

The gpt-4o model family (both gpt-4o and gpt-4o-mini) will not deploy on the UAT cognitive account, while gpt-4.1-mini deploys without issue on the same account. The identical configuration works on the DEV cognitive account in the same subscription and region. This suggests there is something specific about how the UAT account is registered internally on Azure's side that is blocking the gpt-4o model family.

Has anyone seen this before, or can a Microsoft engineer investigate what error 715-123420 indicates for this account?