Changing ChatGPT 5.2 Deployment to Regional or DataZoneStandard

Catalina Castillo Trivino 0 Reputation points
2026-08-12T14:53:24.3366667+00:00

We are currently using ChatGPT 5.2 in Canada to process Protected Health Information (PHI). We need to change our current GlobalStandard deployment to Regional or DataZoneStandard to comply with Quebec’s privacy requirements.

However, we have been unable to complete the configuration because the Azure portal displays an “Insufficient quota” error. This is confusing because the default quota is 1,000,000 tokens per minute, and the documentation indicates that ChatGPT 5.2 requires between 300,000 and 1,000,000 tokens per minute for a DataZoneStandard deployment.

Could you please clarify why we are receiving the “Insufficient quota” error and advise how we can successfully switch to a Regional or DataZoneStandard deployment while keeping our Pay-As-You-Go subscription, without increasing our current costs, so that we can comply with Quebec’s privacy requirements?

Any guidance would be greatly appreciated.

Azure OpenAI in Foundry Models

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. SRILAKSHMI C 19,730 Reputation points Microsoft External Staff Moderator
    2026-09-02T15:57:58.01+00:00

    Hello @Catalina Castillo Trivino

    Thank you for reaching out to Microsoft Q&A.

    The “Insufficient quota” message can occur even when your existing GPT-5.2 Global Standard deployment shows 1,000,000 TPM. Azure OpenAI quota is allocated separately based on the subscription, region, model, and deployment type. Therefore, the quota available for your existing Global Standard deployment does not automatically become available for a Data Zone Standard or Regional Standard deployment.

    For GPT-5.2, the currently documented quota is:

    • Global Standard: 1,000,000 TPM

    Data Zone Standard: 300,000 TPM

    Therefore, the 1,000,000 TPM associated with Global Standard should not be interpreted as 1,000,000 TPM being available for GPT-5.2 Data Zone Standard.

    Recommended approach

    1. Verify deployment availability In Microsoft Foundry, go to Models + endpoints → Deploy model and check whether GPT-5.2 is available with the required Standard/Regional or Data Zone Standard deployment type in the target Canadian region. Model availability varies by region and deployment type, so the availability table should be used to confirm the supported combination.
    2. Check the quota specifically for the target deployment type In Management Center → Quota, select the relevant subscription and region and verify the quota for GPT-5.2 Data Zone Standard (or the applicable Regional Standard option). If the required quota is already allocated, you can free unused TPM from existing deployments where applicable. Microsoft documents that quota is consumed as TPM is assigned to deployments.
    3. Request additional quota if required If there is insufficient quota for the required deployment type and region, submit a quota request through the Foundry quota management experience. Microsoft recommends requesting quota for the specific model and target region/deployment type.
    4. Deploy using the required deployment type Once the required quota is available, create the GPT-5.2 deployment using the supported Regional/Standard or Data Zone Standard option. Please note that the deployment type itself determines the processing boundary: Standard/Regional processes inference in the deployment region, whereas Data Zone deployments process inference within the applicable Microsoft-defined data zone.

    Regarding Pay-As-You-Go and cost

    You can continue using a Pay-As-You-Go subscription; quota allocation and billing are separate concepts. Quota represents the TPM rate limit allocated to the deployment and is not itself a charge for consuming that TPM.

    However, we cannot guarantee that the overall cost will remain exactly the same after changing the deployment type. Pricing depends on the model and deployment type, as well as your actual token consumption. The quota requirement should therefore be considered separately from the billing impact.

    Finally, because your requirement is specifically related to Quebec PHI/privacy compliance, we recommend confirming that the selected deployment type and processing boundary satisfy your organization's specific regulatory requirements before moving the workload.

    please review the relevant Microsoft documentation here:

    Manage Azure OpenAI quota in Microsoft Foundry

    Azure OpenAI quotas and limits

    Model region availability by deployment type

    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. AI answer

    2026-08-12T14:53:46+00:00

    The most likely cause is that quota is assigned per subscription, region, model, and deployment type in TPM, and the available quota for GlobalStandard does not automatically mean the same quota exists for DataZoneStandard in Canada. For gpt-5.2, the documented Tier 1 limits are:

    • GlobalStandard: 1,000,000 TPM
    • DataZoneStandard: 300,000 TPM

    That means a deployment can still show Insufficient quota if the subscription has no available TPM allocated for gpt-5.2 DataZoneStandard in the target region, even when gpt-5.2 GlobalStandard quota exists.

    A second documented factor is that some models can appear in the catalog or deployment flow but not yet have an initial baseline TPM allocation available for self-service quota management. In that case, the quota UI does not show rows that can be adjusted, and a manual quota request is required before deployment can succeed.

    To move from GlobalStandard to DataZoneStandard while staying on a Pay-As-You-Go subscription:

    1. In Azure AI Foundry, check the quota for the target model + deployment type + region combination, specifically gpt-5.2 DataZoneStandard in the Canada region being targeted.
    2. If no quota row exists for that combination, submit a manual quota increase request using the official Azure OpenAI Service quota request form: https://aka.ms/oai/stuquotarequest
    3. Request only the minimum TPM needed for the new deployment. The documented Tier 1 minimum for gpt-5.2 DataZoneStandard is 300,000 TPM.
    4. After the request is approved and the quota appears in Foundry, create the new DataZoneStandard deployment and assign the approved TPM.

    The context does not provide support for a Regional deployment path for gpt-5.2 in this scenario. It does explicitly state that Global Standard is the default deployment mode and Data Zone Standard is supported as an alternative when data residency is required.

    The context also does not state that switching deployment type changes pricing or guarantees no cost change. It only supports that the deployment can remain on Pay-As-You-Go and that quota must exist for the target deployment type before deployment succeeds.

    References:

    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.