I am unable to deploy model in foundry it says I do not have quota

Anand R 0 Reputation points
2026-07-24T22:45:22.7033333+00:00

I am building a Retrieval-Augmented Generation (RAG) application for a project, using the standard "Azure OpenAI On Your Data" architecture (Azure Blob Storage → Azure AI Search → Azure OpenAI chat model). My subscription is upgraded to Pay-As-You-Go and I am prepared to pay standard usage rates.

I currently have 0 quota for all current Azure OpenAI models under the Global Standard deployment type in East US 2, which prevents me from deploying any model. The only models with existing quota (e.g. gpt-4o-mini 2024-07-18) are deprecated and can no longer be deployed, and all current models are Global-Standard-only. I am therefore blocked from deploying any usable chat or embedding model.

I am requesting a modest quota (30K TPM each) sufficient to deploy one chat model and one embedding model for a low-volume, single-user academic demo. This is not a production or high-throughput workload. I would greatly appreciate approval so I can complete the project.

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. Pavan Kumar Purilla 90 Reputation points Microsoft External Staff Moderator
    2026-07-27T11:14:48.4+00:00

    Hi Anand R,
    Thank you for reaching out to Microsoft Q&A.

    From your description, it appears that your Azure OpenAI resource has 0 TPM (Tokens Per Minute) quota available for the current Global Standard models in the East US 2 region. Since Global Standard deployments require available quota, you won't be able to deploy the model until quota is assigned.

    As your subscription is already Pay-As-You-Go, the next step is to request a quota increase for the required models. Since your workload is a low-volume academic RAG application, you can request a modest quota (for example, 30K TPM for one chat model and 30K TPM for one embedding model). Please note that quota requests are reviewed based on service capacity and eligibility, so approval is not guaranteed.

    You can submit a quota request and learn more here:

    If the required quota is unavailable in East US 2, you may also consider deploying the models in another supported region where quota is available.

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

    Was this answer helpful?

    0 comments
  2. kagiyama yutaka 4,920 Reputation points
    2026-07-25T06:03:47.15+00:00

    I think u just check Foundry→Quota, and when all Global Standard models show 0 TPM in a region, that’s simply no quota assigned for that model‑type and deployment‑type there. Current models need Global Standard quota, so deploy stays blocked until quota is approved. You can send the Azure OpenAI quota‑request, and you can use a region that already has Global Standard quota. Deprecated models may still show quota from older rules, but they can’t be deployed anymore.

    Was this answer helpful?

    0 comments

  3. AI answer

    2026-07-24T22:45:43.1633333+00:00

    Quota for Azure OpenAI in Foundry is assigned to the subscription on a per-region, per-model, and for Foundry quota also per-deployment-type basis, in Tokens-per-Minute (TPM). If quota is 0 for the model and deployment type in East US 2, deployment is blocked until quota is available.

    Supported actions are:

    1. In Azure AI Foundry portal, go to Management > Quota to view quota allocations for the target region and model.
    2. If existing deployments are consuming TPM, reduce TPM on those deployments to free quota for a new deployment of the same model.
    3. If no usable quota exists, submit the quota increase request form: https://aka.ms/oai/stuquotarequest.
    4. If East US 2 has no supported quota/capacity for the required model and deployment type, deploy in a supported region where quota is available.

    Useful checks:

    • In Foundry, Management > Quota shows quota allocation and total approved quota for each model type in a region.
    • When creating a deployment, TPM is assigned in increments of 1,000.
    • TPM allocation controls deployment rate limits; RPM is derived from TPM and varies by model.
    • Quota allocation is not related to a model’s maximum input token limit.

    For programmatic verification, the Model Capacities API can be used to check where capacity is available for a model across locations and deployment types in the subscription.

    If the required current models show 0 quota under Global Standard in East US 2, the practical resolution is to request quota or use another supported region with available quota/capacity.

    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.