Unable to deploy any base model from Azure AI foundry due to insufficient quota error

Saurabh 0 Reputation points
2026-10-27T00:41:14.59+00:00

I am unable to deploy any base model from Azure AI foundry as I see insufficient quota error. I requested quota increase multiple times from multiple regions but they all have been rejected! What should I do?

User's image

Foundry Models
Foundry Models

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

0 comments

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Senthil kumar 2,995 Reputation points
    2026-10-27T05:29:52.4166667+00:00

    Hi @Saurabh

    Please check below points and give your feedback.

    1. Verify the Quota page

    In Azure AI Foundry:

    Management → Quotas

    Check:

    • Model name (GPT-6.1 Sol)
    • Scope (Global Standard / Data Zone / Region)
    • Current quota
    • Available quota

    If it shows 0 assigned quota, deployment will always fail. [learn.microsoft.com]

    2. Check your subscription type

    Quota approvals are much more restrictive on:

    • Free Trial
    • Azure for Students
    • Some Sponsorship subscriptions
    • Newly created subscriptions

    If you're using one of these, Microsoft may reject requests due to capacity policies.

    3. Try a different deployment type

    When deploying, check whether:

    • Global Standard
    • Data Zone Standard
    • Standard

    are available.

    Some customers can deploy successfully with a different deployment option even when a specific SKU has no available quota. The deployment menu in your screenshot suggests multiple deployment modes may be available.

    4. Check whether other models have quota

    Look for:

    • GPT-4.1
    • GPT-4o
    • GPT-4o-mini
    • GPT-5-mini (if available in your subscription)
    • Other Foundry-hosted models

    Sometimes Microsoft allocates quota for certain models but not for the newest reasoning models.

    5. Open an Azure Support ticket

    If quota requests have already been rejected multiple times, this is usually the only escalation path.

    Provide:

    • Subscription ID
    • Azure AI Foundry Project name
    • Model name
    • Region
    • Screenshot of the quota page
    • Screenshot of the rejection email

    Thanks.

    Was this answer helpful?

    0 comments

  2. AI answer

    2026-10-27T00:42:13.4966667+00:00

    Use Foundry’s shared quota or Instant access for evaluation; if a dedicated deployment is required, select a model/deployment type with existing quota or open an Azure support request because repeatedly submitting the rejected quota form will not create deployable capacity.

    1. Check whether usable quota already exists

    In Foundry (new):

    1. Open the correct project and subscription.
    2. Select Manage → Quota → Token per minute.
    3. Locate the required model/version and inspect:
      • Scope
      • Region
      • Deployment type
      • Allocated and available TPM
      • Affiliated deployments using shared quota
    4. If another deployment has unused TPM, select its pencil action and reduce that allocation. Allow up to 15 minutes for the released quota to appear.

    Quota is allocated by subscription, model, region or quota scope, and deployment type. Quota assigned to one combination cannot necessarily be used by another. Viewing it requires Cognitive Services Usages Reader at subscription scope; editing allocations requires Cognitive Services Contributor plus that reader role.^1^

    The screenshot’s disabled Default settings means that particular model’s default deployment option has no usable quota. Open Deploy → Custom settings and inspect the available deployment types. If offered and your data-processing requirements permit it, try Global Standard, which has the broadest region coverage and highest default quota. Changing deployment type only helps when that type has a nonzero quota pool; it does not bypass quota.^2^

    2. Unblock evaluation without dedicated quota

    Two documented alternatives may be available:

    • Shared quota: From the model catalog, create a temporary test endpoint using shared quota when that option appears. Shared quota is usage-billed and intended only for temporary testing, not production.^3^
    • Instant access (preview): If the model supports it, invoke it without creating a deployment. In a West US 3 project, open Discover → Models and filter Deployment options → Instant. Instant access uses a separate per-model global quota pool, although the subscription must still have quota for that model.^4^

    If neither option appears, choose another model for which Manage → Quota shows available TPM. The catalog displaying a model does not itself establish that the subscription has deployable quota.

    3. Escalate the rejected requests

    Since multiple quota requests have already been rejected, do not submit duplicate forms. Quota requests are prioritized for subscriptions actively consuming their existing allocation, and requests that do not meet that condition can be denied.^5^

    Create a formal case instead:

    1. In the Azure portal, open Help + support → Create a support request.
    2. Select Service and subscription limits (quotas) and the Azure OpenAI/Foundry quota issue.
    3. Include:
      • Subscription type
      • Exact model and version
      • Deployment type, preferably Global Standard if acceptable
      • Requested TPM
      • Regions already attempted
      • Rejected request identifiers
      • Business use case, expected traffic, and target date
      • Screenshot showing zero/insufficient quota
    4. Ask support to confirm whether the subscription is eligible for any dedicated quota and which model/region/deployment-type combination currently has capacity.

    Subscription limitation to check

    If this is an Azure for Students subscription, modern interactive models can have zero default quota and approval is not guaranteed. A similar student-subscription case was resolved by moving to Pay-As-You-Go, Azure sponsorship, or an education subscription with fewer offer-level restrictions.^6^

    That conclusion should only be applied after checking Azure portal → Subscriptions → your subscription → Overview → Offer/subscription type. For a normal Pay-As-You-Go, CSP, or enterprise subscription, use the support escalation above rather than changing subscriptions immediately.

    References

    1. Manage Azure OpenAI in Microsoft Foundry Models quota
    2. Deployment types for Microsoft Foundry Models
    3. Manage and increase quotas for resources with Microsoft Foundry (Foundry projects)
    4. Instant access to models in Microsoft Foundry (preview)
    5. Microsoft Foundry Models quotas and limits
    6. Quota Issues (0 TPM) and Regional Restrictions on "Azure for Students" Subscription for AI Foundry - Microsoft Q&A
    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.