How to Request Quota for Codex & Claude Models in Microsoft Foundry

The Problem

You’ve found the perfect model in the Microsoft Foundry catalog — maybe GPT-5.3-Codex for agentic coding or Claude Opus 4.6 for complex reasoning. You hit “Deploy” and… “No quota available.” Welcome to one of the most frustrating experiences in Foundry right now. The quota system for non-OpenAI-standard models works differently than what most IT Pros expect, and it’s blocking teams from even evaluating frontier models.

The Nugget

Codex models (gpt-5.x-codex) require a separate registration for limited access, while Claude (Anthropic) models have their own isolated quota pool — neither shares quota with your standard Azure OpenAI deployments. You need to request access and quota through different paths depending on the model provider.

How to Get Access

Step 1: Understand the Access Model

Model FamilyAccess TypeQuota PoolRegistrationGPT-5.x-Codex (gpt-5.3-codex, gpt-5.2-codex, etc.)Limited AccessAzure OpenAI quota (separate from standard models)Required — Apply hereClaude (Opus 4.6, Sonnet 4.6, Haiku 4.5)MarketplaceAnthropic-specific quota (completely separate)Azure Marketplace terms requiredStandard Azure OpenAI (GPT-4.1, GPT-5-mini, etc.)Generally AvailableStandard Azure OpenAI quotaNone

Step 2: Request Model Access

For Codex models: Registration is required. If you’ve previously received access to any limited access model (like gpt-5 or o3), your subscription is auto-approved for new Codex releases. If not, apply through the Azure OpenAI limited access form.

For Claude models: Accept the Anthropic Marketplace terms in Azure Marketplace first. Quota is not auto-assigned — many subscriptions start at 0/0 and need explicit backend enablement.

Step 3: Request Quota Increase

Go to Foundry Portal → Management Center → Quota to check your current allocation. If it shows 0/0:

Submit the Quota Increase Request Form for Azure OpenAI & Anthropic models
For Claude specifically, open an Azure Support ticket requesting “Global Standard enablement” — the form alone may not be enough
Include your subscription ID, target model, region, and use case justification

Why This Matters

Codex and Claude quotas are isolated — deleting other model deployments won’t free up capacity for them
Registration ≠ Quota — having model access approved doesn’t mean you have deployment quota allocated
Priority goes to active users — Microsoft prioritizes quota increases for subscriptions that actively consume their existing allocation
Region matters — Claude models are currently available in East US 2 and Sweden Central only via Global Standard deployment

💡 Pro Tip: If you’re stuck at 0/0 quota for Claude models, don’t rely on the quota form alone. Open a formal Azure Support ticket (Help + Support → Technical → Cognitive Services) and specifically request “initial Anthropic Claude quota allocation.” The standard form creates a catch-22 for subscriptions starting from zero. EA/MCA-E customers should mention their agreement type — it accelerates approval significantly.

Try It Now

Check your current quota: Foundry Portal → Operate → Quota
Apply for Codex access: aka.ms/oai/access
Request quota increase: aka.ms/oai/stuquotarequest
Read the full docs: Foundry Models Quotas & Limits

Runbook notes for How to Request Quota for Codex & Claude Models in Microsoft Foundry

How to Request Quota for Codex & Claude Models in Microsoft Foundry deserves a little more operational context because the decision usually affects AI quota request. The related items are Codex quota, Claude quota, Microsoft Foundry, model capacity, approval workflow. Treat this Nugget as a starting point for a concrete tenant decision: who is in scope, which Microsoft portal or policy is touched, and what visible result should confirm that the configuration worked.

When validating How to Request Quota for Codex & Claude Models in Microsoft Foundry, keep the test narrow enough to understand the result. Select one representative user, device, workload or subscription, capture the current state, then apply the change and compare the outcome. This avoids guessing later when support sees a different enrollment state, access result, model response, update status or admin center signal.

The most useful documentation for How to Request Quota for Codex & Claude Models in Microsoft Foundry is practical rather than theoretical. Record the assignment logic, the owner, the expected monitoring view and the exception path. If the change affects users, include the wording support teams should use when they explain the behavior. If it affects devices or services, include the exact place where administrators can verify health.