Microsoft Foundry shows 0/0 TPM quota for multiple advanced models

Aill same 0 Reputation points
2026-08-22T17:24:42.42+00:00

Hello,

I recently converted my Azure subscription from the $200 introductory credit/trial to Pay-As-You-Go, and my subscription is active.

In Microsoft Foundry → Operate → Quota, I noticed that many advanced models from different providers show 0/0 TPM. This is not limited to Azure OpenAI models.

For example, advanced models from providers such as Anthropic and other model providers also show 0/0 TPM, which prevents me from creating deployments.

I understand that quota and capacity can vary by region and model. However, because multiple models from different providers are showing 0/0 TPM, I would like to understand whether this is expected for a newly converted Pay-As-You-Go subscription.

Could someone please clarify:

  1. Is 0/0 TPM for multiple models normal for a new Pay-As-You-Go subscription?
  2. Does the subscription need to receive an initial quota/entitlement before these models can be deployed?
  3. Is there any activation or provisioning step required after converting from the introductory credit to Pay-As-You-Go?
  4. If this is expected, what is the correct way to obtain quota for these models?

I can provide screenshots of the Microsoft Foundry Quota page showing the 0/0 TPM values.

Thank you.

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

3 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. SRILAKSHMI C 19,730 Reputation points Microsoft External Staff Moderator
    2026-09-01T15:44:35.2733333+00:00

    Hello @Aill same

    Thank you for reaching out to Microsoft Q&A.

    Seeing 0/0 TPM for multiple models in Microsoft Foundry does not necessarily indicate an issue with the Pay-As-You-Go conversion. Quota and deployment eligibility are evaluated based on the subscription, model, deployment type, and region, and partner models can have additional eligibility requirements.

    For Azure-sold models, quota is assigned at the subscription level and can vary by model and deployment type. If the available quota for an eligible model is 0, deployment cannot proceed until quota is available, either through an existing allocation/reallocation or an approved quota increase. Microsoft documents that quota increases can be requested for Foundry Models sold by Azure, Azure OpenAI models, and Anthropic models.

    For partner/community models such as Anthropic, there are additional requirements. These models use Azure Marketplace, and availability depends on the model, subscription type, billing country/region, and deployment region. In particular, Anthropic requires a paid Azure subscription with an active Pay-As-You-Go billing method; free-trial, student, and other unsupported credit-based subscriptions are not eligible.

    Therefore, converting the subscription from the introductory $200 credit to Pay-As-You-Go does not mean that every model will automatically receive non-zero quota. It primarily satisfies the paid-subscription requirement for models that require it. There is no general additional "activation" step documented for Pay-As-You-Go conversion.

    Please check

    Confirm the subscription has an active Pay-As-You-Go billing method.

    Check the specific model and deployment type that shows 0/0 TPM.

    Verify that the model is available in the selected Azure region and that the subscription is eligible for that model.

    For partner models such as Anthropic, verify the required Azure Marketplace subscription/permissions and provider-specific eligibility.

    If the model is eligible and supports quota increases, submit a quota increase request. Microsoft notes that quota increase requests are supported for Azure-sold Foundry Models, Azure OpenAI models, and Anthropic models. However, not all partner/community models support quota increases.

    For example, the current Claude documentation shows that Pay-As-You-Go subscriptions have defined rate limits for supported Claude models, while some model/deployment combinations are marked N/A rather than having an allocatable quota. This demonstrates why the exact model and deployment type need to be checked rather than treating all 0/0 entries as a single quota issue.

    Please refer this

    Claude models in Microsoft Foundry – quotas and subscription eligibility

    Deploy and use Claude models in Microsoft Foundry

    Microsoft Foundry Models quotas and limits: https://learn.microsoft.com/azure/foundry/foundry-models/quotas-limits?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#quotas-and-limits-reference

    Manage Azure OpenAI in Microsoft Foundry Models quota (model-specific settings): https://learn.microsoft.com/azure/foundry/openai/how-to/quota?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#model-specific-settings

    Deploy Microsoft Foundry Models in the Foundry portal (quota for deploying and running inference + troubleshooting): https://learn.microsoft.com/azure/foundry/foundry-models/how-to/deploy-foundry-models?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#quota-for-deploying-and-running-inference-on-a-model

    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?

    1 person found this answer helpful.
    0 comments
  2. Megha Ramakrishnan 500 Reputation points
    2026-08-23T10:52:14.8233333+00:00

    Hi @Aill same

    Welcome to Microsoft Q&A.

    0/0 TPM quota for multiple advanced models (including Azure OpenAI, Anthropic, and others) after converting from the $200 trial to Pay‑As‑You‑Go — is expected behavior for a newly converted subscription.

    Why this happens:

    • Trial vs. Pay‑As‑You‑Go: The $200 introductory credit subscription includes limited default entitlements. When converted to Pay‑As‑You‑Go, those entitlements do not automatically carry over.

    • Quota assignment: Advanced models require quota approval. By default, a new Pay‑As‑You‑Go subscription shows 0/0 TPM until quota is explicitly granted.

    Steps needed to be performed:

    1. No activation step required

    • Converting to Pay‑As‑You‑Go activates the subscription, but it does not automatically assign quota for advanced models.

    1. Request quota

    • In Microsoft Foundry → Operate → Quota, use the Request quota option for the models you want to deploy.

    • Quota requests are reviewed and approved based on capacity, region, and responsible AI usage guidelines.

    1. Wait for approval

    • Once approved, the quota will update from 0/0 TPM to the assigned throughput per minute.

    • Only then will you be able to create deployments for those models.

    References:

    Request form for quota increase: https://aka.ms/oai/stuquotarequest

    https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/quota?tabs=python

    Help make this community better for everyone: if this answer resolved your issue, please accept it or leave an upvote. If not, share more details in a comment so we can continue the discussion and find the right solution

    Was this answer helpful?

    1 person found this answer helpful.
    0 comments

  3. AI answer

    2026-08-22T17:25:03.5+00:00

    Yes. This can be expected for a newly converted or newly active Pay-As-You-Go subscription.

    Quota in Foundry is assigned at the subscription level and is scoped per region, per model, and per deployment type in Tokens-per-Minute (TPM). A model can appear in the catalog but still show no usable quota in Operate > Quota or show 0/0 TPM, which means there is no initial baseline allocation/entitlement available yet for that subscription-model-region combination. In that state, deployment is blocked.

    For the scenarios described:

    1. Is 0/0 TPM normal for multiple models on a new Pay-As-You-Go subscription?
      Yes. The documented quota model is per subscription, per region, and per model. The accepted Microsoft Q&A guidance also states that when a model appears in the catalog but shows insufficient quota and no quota row, it typically means the subscription does not yet have a default TPM allocation for that model.
    2. Does the subscription need an initial quota/entitlement first?
      Yes. A deployment requires available quota for that model. If no baseline allocation exists, the model cannot be deployed until quota is granted.
    3. Is there an activation or provisioning step after converting from introductory credit to Pay-As-You-Go?
      No separate activation step is documented in the provided material. The supported path described is quota assignment and quota request management, not a post-conversion activation workflow.
    4. What is the correct way to obtain quota?
      Use one of these supported paths:
      • In Foundry, go to Manage > Quota.
      • Review the model deployment entry and current allocation.
      • If quota exists elsewhere, reallocate TPM from affiliated or other deployments to free capacity.
      • If no usable quota exists, select Request quota in the upper-right corner of the Quota page, or submit the quota increase request form: https://aka.ms/oai/stuquotarequest

    Additional points supported by the documentation:

    • Quota changes can take up to 15 minutes to propagate after editing allocation or submitting a request.
    • Quota increase requests are processed in received order, with priority for customers actively using existing quota.
    • Quota increases are supported for Foundry Models sold by Azure, Azure OpenAI models, and Anthropic models. Except for Anthropic models, models from partners and community do not support quota increases.

    So for multiple providers showing 0/0 TPM, the practical interpretation is:

    • the subscription currently has no deployable quota for those model/region/deployment-type combinations, and
    • the next action is to check Manage > Quota, then request quota where supported, or reallocate existing TPM if any is already assigned.
    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.