An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
Yes. Quota availability by region for a subscription can be checked in the Azure AI Foundry portal at Azure AI Foundry portal.
For a specific model and version, regional capacity can also be checked programmatically with the capacity API. The API requires subscriptionId, model_name, and model_version, and returns available capacity for that model across regions and deployment types for that subscription.
Key quota behavior:
- Quota is assigned per subscription, per region, and per model in TPM.
- New deployments consume TPM from that regional model quota.
- If quota is exhausted, testing can only continue by reducing TPM on other deployments or requesting a quota increase.
For quota increases, submit the quota increase request form. Requests are processed in order received, and priority is given to customers already consuming their existing quota allocation.
One important limitation: the portal and capacity API can return quota/capacity information for retired models that are no longer available for new deployments.
References: