How to know my quota limit for specific Foundry models?

Ritabrata Chakraborty 0 Reputation points
2026-08-04T06:14:30.0433333+00:00

I received $10,000 in Azure credits through YC Startup School / Microsoft for Startups. I would like to confirm my Azure OpenAI access.

Specifically, I would like to know:

Whether my subscription is eligible to use Azure OpenAI.

Which OpenAI models are available (e.g., GPT-5, GPT-5 mini, GPT-4.1, o3, embeddings).

The default TPM (Tokens Per Minute) and RPM (Requests Per Minute) quotas for each available model.

Whether I need to submit a quota request or complete any additional steps before deploying Azure OpenAI models.

Thank you.

Azure OpenAI in Foundry Models
0 comments

3 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. SRILAKSHMI C 19,730 Reputation points Microsoft External Staff Moderator
    2026-08-04T16:45:39.8533333+00:00

    Hello @Ritabrata Chakraborty

    Thank you for reaching out to Microsoft Q&A.

    If you've received Azure credits through Microsoft for Startups (including the YC Startup School program) and would like to confirm your Azure OpenAI access, you can verify your subscription eligibility, available models, and assigned quota directly from the Azure AI Foundry portal.

    1. Verify whether your subscription is eligible to use Azure OpenAI

    Sign in to Azure AI Foundry: https://ai.azure.com/

    Navigate to Management Center → Quota.

    From there, select your:

    • Subscription
    • Region
    • Model

    If you can view quota information for a model in your selected region, your subscription has access to that model. If the allocated quota is 0 TPM, or the model does not appear, it may indicate that the model is not currently enabled for your subscription or region, or that additional approval is required before deployment.

    2. Check which OpenAI models are available

    Model availability depends on your subscription, Azure region, and deployment type (for example, Global Standard or Provisioned).

    3. View your TPM and RPM quota

    Azure does not publish a single default TPM/RPM value that applies to every customer. Quotas are assigned based on the subscription, model, region, and deployment type.

    To view your assigned quota:

    1. Go to Azure AI Foundry → Management Center → Quota.
    2. Select your Subscription, Model, and Region.
    3. Review the Allocated TPM and Used TPM values for each model.

    The quota page provides the authoritative quota assigned to your subscription. Corresponding request rate limits (RPM) are derived from the allocated TPM and are enforced by the service.

    4. Do I need to request additional quota before deploying?

    If the assigned quota is sufficient for your workload, you can deploy supported models immediately.

    If the allocated quota is 0 TPM or lower than your workload requires, you can request a quota increase directly from the Quota page in Azure AI Foundry.

    Quota requests are reviewed based on factors such as your subscription type, requested model, Azure region, and available service capacity.

    In some cases, if deployment fails with an entitlement-related error (for example, SpecialFeatureOrQuotaIdRequired), the model may require additional approval or limited-access registration before it can be deployed.

    Recommended next steps

    • Sign in to Azure AI Foundry (https://ai.azure.com/).
    • Navigate to Management Center → Quota.
    • Select your Subscription, Region, and the models you're interested in (for example, GPT-5, GPT-5 mini, GPT-4.1, o3, embeddings).
    • Verify whether each model has an Allocated TPM greater than zero.
    • If additional quota is required, submit a quota request from the Quota page.

    Please refer this

    I Hope this helps. Do let me know if you have any further queries.

    If this answers your query, please do click Accept Answer and Yes for was this answer helpful.

    Thank you!

    Was this answer helpful?

    0 comments
  2. Gowtham CP 8,120 Reputation points Volunteer Moderator
    2026-08-04T06:18:54.45+00:00

    Hi @Ritabrata Chakraborty ,

    Thanks for your question.

    Your $10,000 Microsoft for Startups / YC Startup School credits do not determine your Azure OpenAI quotas or model access. These depend on your subscription, region, and assigned quota tier.

    You can verify everything from Azure AI Foundry:

    • Model Catalog – Shows which OpenAI models (GPT-5, GPT-4.1, o3, embeddings, etc.) are available for your subscription and region.
    • Management → Quotas – Displays your assigned TPM, RPM, and available quota for each model.

    In most cases, you can deploy and use the models with your assigned quota immediately. You only need to submit a quota increase request if your workload exceeds the default allocation.

    References:

    I hope this helps. If it answers your question, please consider accepting it to help others with the same query.

    Was this answer helpful?

    0 comments

  3. AI answer

    2026-08-04T06:14:58.7333333+00:00

    Azure OpenAI quota is assigned to a subscription on a per-region, per-model basis in Tokens Per Minute (TPM). In Microsoft Foundry, quota is used to control deployment rate limits, and the available quota is reduced as TPM is assigned to deployments.

    What can be confirmed:

    1. Whether the subscription is eligible to use Azure OpenAI
      • Azure OpenAI access is not determined by Azure credits alone.
      • A subscription that is onboarded to Azure OpenAI receives default quota for most available models.
      • Eligibility and actual available capacity can be checked in Foundry by opening Operate > Quota in Foundry (new), or Management > Quota in Azure AI Foundry portal.
    2. Which models are available
      • The provided quota and limits context includes these model families or models:
        • gpt-4.1
        • gpt-4.1-mini
        • gpt-4o
        • gpt-4o-mini
        • o3
        • o3-mini
        • o3-pro
        • o4-mini
        • o1
        • o1-preview
        • o1-mini
        • text-embedding-3-large
        • text-embedding-3-small
        • text-embedding-ada-002
        • gpt-5
        • gpt-5.1
        • gpt-5.4
        • gpt-5.4-mini
        • gpt-5.5
        • gpt-oss-120b
      • Actual availability depends on the subscription, region, and deployment type.
    3. Default TPM and RPM quotas from the provided context
      • For these model capacity units:
        • Older chat models: 1 unit = 1,000 TPM and 6 RPM
        • o1 & o1-preview: 1 unit = 6,000 TPM and 1 RPM
        • o3: 1 unit = 1,000 TPM and 1 RPM
        • o4-mini: 1 unit = 1,000 TPM and 1 RPM
        • o3-mini: 1 unit = 10,000 TPM and 1 RPM
        • o1-mini: 1 unit = 10,000 TPM and 1 RPM
        • o3-pro: 1 unit = 10,000 TPM and 1 RPM
      • For gpt-oss-120b, the listed limit is 5 M TPM and 5 K RPM.
      • For Azure Government regional standard TPM defaults, the context lists:
        • gpt4.1: 300,000 TPM default
        • gpt-4.1-mini: 300,000 TPM default
        • gpt-4o: 80,000 TPM default
        • text-embedding-3-large: 350,000 TPM default
        • text-embedding-3-small: 350,000 TPM default
        • text-embedding-ada-002: 350,000 TPM default
      • The provided context does not include a complete default TPM/RPM table for all listed commercial Foundry models such as GPT-5, GPT-5 mini, GPT-4.1, or embeddings outside the specific entries above.
    4. Whether a quota request or additional steps are needed before deployment
      • If enough quota is already available for the target model and region, deployment can proceed by assigning TPM during deployment creation.
      • If quota is exhausted, a new deployment can only be created by:
        • reducing TPM assigned to other deployments of the same model, or
        • requesting a quota increase in the desired region.
      • In Foundry (new), quota can be viewed and managed at Operate > Quota. The Request quota button can be used to request increases for standard deployments.
      • After editing quota allocation or submitting a request, allow up to 15 minutes for changes to propagate, then refresh the Quota page.
      • Quota increase requests can be submitted through the quota increase request form.

    To verify the current quota for a specific subscription and model in Foundry:

    1. Sign in to Microsoft Foundry.
    2. Turn on New Foundry.
    3. Open the target project.
    4. Select Operate.
    5. Select Quota.
    6. Review the Token per minute tab for standard deployments.
    7. Select a deployment to see its current allocation, usage, and affiliated deployments.
    8. Use Request quota if the required model or region does not have enough available quota.
    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.