How to fix "Failed to deploy model GPT-5.4-mini our system has detected this request as unusual activity for your account. Contact support"

Question

I tried deploying GPT-5.4-mini model on Azure AI foundry and received this error "Failed to deploy model GPT-5.4-mini our system has detected this request as unusual activity for your account. If you think this is a mistake then Contact support".
Need help.
Thank You
Attaching the screenshot:

Answer 1

Hello @Abhay Gupta ,

Welcome to Microsoft Q&A .Thank you for reaching out to us.

The deployment failure message is generally associated with an automated security validation process that evaluates activity patterns before allowing certain model operations. This validation helps protect Azure AI services from unauthorized or unusual usage patterns. In some cases, legitimate deployment attempts may also require additional verification before proceeding.

It needs respective backend teams to further review and assist on this case.

The following checks can help validate the environment and address the deployment issue:

1. Validating subscription and resource health • Confirm that the Azure subscription is active and does not have any billing, compliance, or service restrictions. • Verify that the Azure AI Foundry project and associated AI resources are successfully provisioned. • Review recent deployment events and failures in Azure Activity Logs for:
   • Operation details
   • Failure information
   • Correlation identifiers
   • Related deployment activities
2. Confirming model availability and deployment eligibility • Verify that GPT-5.4-mini is available in the selected Azure region. • Confirm that the AI resource is eligible to deploy the requested model. Verify that quota is available for:
   • The model
   • The deployment type
   • The selected region
   Attempt the deployment directly through the Azure AI Foundry portal to rule out issues related to:
   • Automation scripts
   • CLI-based deployments
   • API-based deployment requests
3. Validating that the deployment identity has the required Azure role assignments. • Confirm permissions at:
   • Subscription scope
   • Resource Group scope
   • Azure AI resource scope
   Common roles used for deployment operations include:
   • Contributor
   • Cognitive Services Contributor
4. Determining the scope of impact
5. Attempt deployment of another supported model. • Verify whether the issue affects:
   • Only GPT-5.4-mini
   • Multiple model deployments
   • Confirm whether existing deployments continue to function successfully. This validation helps determine whether the behavior is model-specific or related to a broader subscription-level deployment restriction.

The following references might be helpful , please check them out

Activity Log in Azure Monitor - Azure Monitor | Microsoft Learn

What is Azure role-based access control (Azure RBAC)? | Microsoft Learn

We have reached out to you on private chat for further assistance

Please let us know if the response was helpful

 

Thank you

Answer 2

The wording of this error matters: unusual activity for your account is not the normal quota/capacity error. It indicates that the deployment request was stopped during a service-side account or subscription validation check, before model capacity was allocated. Increasing TPM or repeatedly changing the deployment name is therefore unlikely to fix it.

Check these once:

• the subscription is Active, billing is valid, and you are in the intended tenant;
• Microsoft.CognitiveServices is registered and your role can create account deployments;
• GPT-5.4-mini is shown as deployable in the selected region and the matching deployment type has available quota under Foundry > Operate > Quota;
• Activity Log contains the failed operation's exact UTC time plus any trace, correlation, client-request, or service-request ID.

If those checks pass, open a case from the Foundry/Cognitive Services resource via Support + troubleshooting > Create a support request. Select Cognitive Services / Azure OpenAI and ask for an account/subscription-level deployment-block review. Provide the subscription, tenant, resource, region, model version/SKU, screenshot, timestamp, and request IDs privately. Also state whether existing deployments still work.

There is no portal setting to clear this flag. Avoid repeated deployment attempts or deleting the resource while it is being reviewed.

References: Foundry support options and Azure OpenAI quota management.