Source: Microsoft Foundry — Azure OpenAI Quotas and Limits

Foundry's quota model: subscription-level pools, automatic quota tiers (Free + Tiers 1–6), TPM/RPM per model and deployment type, and hard service limits.

Contents

Summary

Reference tables for TPM/RPM by tier and model, plus resource-level limits.

Key claims

  • Quota is scoped no higher than the subscription. After May 7, 2026 quota is moving to subscription-level pools, starting with Realtime Translate and Realtime Whisper and later all models: Global Standard shares one pool across regions per model/version, Data Zone Standard one pool per zone.
  • Seven quota tiers (Free/Tier 0, 1–6; Free can’t request increases) are assigned by usage, Microsoft relationship (EA/MCA-E) and payment history, and upgrade automatically unless you opt out (NoAutoUpgrade).
  • RPM is tied to TPM per model version (e.g. 10 RPM or 1 RPM per 1,000 TPM for different gpt-chat-latest versions).
  • Example Tier 1: gpt-4.1 Global Standard 1M TPM / 1,000 RPM; gpt-5 Global Standard 250K TPM.
  • Limits: 30 Azure OpenAI resources per region per subscription, 32 standard deployments per resource, 128 tools per chat request, 2,048 embedding inputs per request, 100,000 PTUs max per deployment.
  • Batch: up to 200 MB per input file (1 GB with bring-your-own storage).

Feeds

Azure AI Foundry

Suggest changes →