Insufficent Quota for all claude base model deployment in us-east-2 and eu-sweden region

Hello Global Admin,

Thank you for reaching out.

The “Insufficient quota” error when deploying Claude (Anthropic) base models in East US 2 and Sweden Central typically indicates that your subscription either does not have quota allocated for these models or the available capacity in that region is currently constrained.

What “Insufficient quota” means

This error generally occurs due to one or more of the following:

• No Tokens Per Minute (TPM) quota assigned to your subscription
• Existing quota already fully consumed
• Regional capacity constraints
• Subscription not eligible for Anthropic models

Steps to resolve

1. Confirm your subscription type

Claude models require supported subscription types such as:

• Enterprise Agreement (EA)
• Microsoft Customer Agreement – Enterprise (MCA-E)
• Pay-as-you-go

Not supported:

• Free trial
• Student / startup credits
• Sponsorship-only subscriptions

If your subscription falls under unsupported types, quota cannot be allocated.

2. Verify roles and permissions

Ensure you have:

• Resource Group level: Owner or Contributor
• Additionally: Cognitive Services User role

If you cannot view quota:

• You may need Cognitive Services Contributor (or equivalent)
• Specifically for permission Microsoft.CognitiveServices/locations/usages/read

3. Check existing quota in Foundry

Navigate to Azure AI Foundry → Management → Model quota

Verify:

• Claude Base model quota is listed
• TPM (Tokens Per Minute) is assigned
• Available quota is not already consumed

If quota is 0 or not listed, you will not be able to deploy models.

4. Free up or reallocate quota

If quota exists but is fully used:

• Reduce TPM assigned to less critical deployments
• Delete unused deployments

Deleted deployments may hold quota for up to 24–48 hours unless fully purged

5. Request a quota increase

To increase quota Go to Azure AI Foundry → Quota → Request increase

Or submit request via:

https://aka.ms/oai/stuquotarequest

Include:

• Subscription ID
• Target region (East US 2 / Sweden Central)
• Model (Claude Base / Sonnet / Opus)
• Requested TPM
• Business justification and expected usage

Requests are reviewed and approved based on:

• Subscription type
• Use case
• Regional capacity

6. Validate model access

Ensure Claude models are visible in Azure AI Foundry → Model Catalog

If not, visible Access is not enabled for your subscription and must be requested separately.

7. Consider regional capacity

Even with quota Deployment can fail if the region is at capacity

Recommended:

• Retry deployment
• Try another supported region

Please refer this

Deploy and use Claude models in Microsoft Foundry (preview) – region & subscription requirements, troubleshooting: https://learn.microsoft.com/azure/foundry/foundry-models/how-to/use-foundry-models-claude?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#prerequisites

Deploy Microsoft Foundry Models in the Foundry portal – quota details & management: https://learn.microsoft.com/azure/foundry/foundry-models/how-to/deploy-foundry-models?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#quota-for-deploying-and-running-inference-on-a-model

Request quota increase: https://aka.ms/oai/stuquotarequest

Cognitive Services Contributor role (for subscription-level usage/read): https://learn.microsoft.com/azure/role-based-access-control/built-in-roles#cognitive-services-contributor

I Hope this helps. Do let me know if you have any further queries.

Thank you!