A catalog of AI models in Microsoft Foundry that you can discover, compare, and deploy using Azure’s built‑in tools for evaluation, fine‑tuning, and inference
Hello @Rizal Muharam ,
Welcome to Microsoft Q&A .Thank you for reaching out to us.
Thank you for sharing the error details.
The behavior described here aligns with a known issue currently under active investigation in the deployment path for Claude / Anthropic models in Azure AI Foundry. Current findings indicate that the provisioning flow may fail while resolving the required provider metadata fields used during deployment particularly organizationName and countryCode which then surfaces the InvalidModelProviderData message seen during deployment attempts. This is being actively worked on by the relevant backend teams.
For Claude model deployments, the required provider metadata includes industry, organizationName, and countryCode. In addition, Claude deployments require a paid Azure subscription with a valid payment method, Marketplace access, and deployment from a supported region such as East US 2 or Sweden Central.
As a precaution, the following validations are recommended:
- Confirm the Foundry resource is in a supported Claude region.
- Confirm the subscription and billing setup are eligible for Claude model deployment.
- If deployment is being attempted through REST API / CLI / automation, ensure that
modelProviderDatais explicitly included with industry, organizationName, and countryCode. - If industry is provided manually, ensure it uses a supported value exactly as expected by the service. Related cases have shown that invalid or differently cased industry values can lead to a separate provisioning failure.
At this stage, however, the current symptom pattern does appear consistent with the ongoing service-side issue rather than a simple misconfiguration in the deployment request. Once the backend remediation is in place, a new deployment attempt should be the next recommended step.
Please let us know if the response was helpful
Thank you
Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the response was helpful. This will be benefitting other community members who face the same issue.