Azure OpenAI models availability in Canada (East/Central) with Standard deployment

Hello @Malek

Thank you for reaching out to Microsoft Q&A.

I understand that you're designing a production solution with the following requirements:

Strict Canadian data residency, where prompts and responses must be processed within Canada.

Preference for Standard (pay-per-token) deployments rather than Provisioned deployments due to cost considerations.

A currently supported GPT model that is neither deprecated nor approaching retirement.

Data residency in Canada

For your data residency requirement, the key point is that Regional Standard deployments process prompts and responses within the Azure region where the Azure OpenAI resource is deployed. Therefore, if your resource is deployed in Canada East or Canada Central using a Regional Standard deployment, request processing is tied to that Canadian region.

By contrast, Global Standard deployments may route requests across Azure's global infrastructure to optimize capacity and throughput, making them unsuitable for strict Canadian data residency requirements.

Model availability

Azure OpenAI model availability depends on three factors:

Azure region

Deployment type (Standard, Global Standard, or Provisioned)

Model/version lifecycle

Not every model is available in every region or under every deployment type. Additionally, model availability changes over time as new versions are introduced and older versions are retired.

Regarding your observations:

• If GPT-4o or GPT-4.1-mini are no longer available for deployment in your Canadian region, this is generally due to the model's lifecycle status or its current regional availability.
• Likewise, if GPT-5 is currently available only as a Regional Provisioned deployment in Canada, this reflects the current published availability for that model and deployment type rather than a configuration issue with your subscription.

Current Standard (Regional) model availability

Based on the current Azure AI Foundry regional availability documentation, Canada East supports several models under Regional Standard deployments, including:

GPT-4o (2024-08-06)

GPT-4o (2024-11-20)

GPT-4o-mini (2024-07-18)

GPT-4.1-mini (2025-04-14)

o1 (2024-12-17)

Embedding models such as:

text-embedding-3-small

  **text-embedding-3-large**
  
     **text-embedding-ada-002**
     

Availability may differ between Canada East and Canada Central, so it's important to verify the specific region you're targeting.

Why GPT-5 or other models may not appear

If GPT-5 (or another model) is not available for a Standard deployment in your Canadian region, common reasons include:

The model is not currently supported in that region for the selected deployment type.

The model requires limited access or additional approval.

The model has entered a "No New Customers" or retirement lifecycle phase for new deployments.

Standard deployment quota is unavailable for that region.

If GPT-5 is currently listed only under Regional Provisioned deployments in Canada, then that is the supported deployment option for the model at this time.

How to verify current availability

Because model availability evolves over time and the Azure AI Foundry portal reflects the latest supported deployments the recommended approach is to verify availability directly in the portal:

1. Navigate to Azure AI Foundry.
2. Select Models + Endpoints > Deploy model.
3. Filter by:
   • Region: Canada East or Canada Central
   • Deployment Type: Standard
4. Review the list of deployable models.

Please verify this

Region availability for Foundry Models sold by Azure (standard) — Standard/Regional availability table: https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability

Azure OpenAI in Azure AI Foundry Models (model summary table and region availability): https://learn.microsoft.com/azure/ai-foundry/openai/concepts/models

Region availability for Foundry Models sold by Azure (standard) — data zone standard (note on processing within zones): https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability

Azure OpenAI models and regions for Foundry Agent Service (classic) (global standard availability table excerpt): https://learn.microsoft.com/azure/foundry-classic/agents/concepts/model-region-support

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!

Hello Malek,

Greetings!

You've actually diagnosed this correctly, and the gap you're seeing is a real platform limitation rather than something wrong on your end. Here is the full picture based on the current official region availability and model retirement data.

Canada Central has no Standard (pay-per-token) model support at all. Looking at the official Foundry region availability table, the Standard/Regional deployment category does not even list Canada Central as a region option. Canada Central only supports Global and Provisioned deployment categories. If strict in-region processing is a requirement, your only Standard deployment option in the Canada geography is Canada East.

In Canada East, the current Standard/Regional model list is narrow. As of today, the chat models supported for Standard deployment in Canada East are:

gpt-4.1-mini (version 2025-04-14)  - GA, retires 2026-10-14
gpt-4o (version 2024-11-20)        - GA, retires 2026-10-01

Both are still GA (generally available and deployable) right now, not deprecated. The earlier gpt-4o versions (2024-05-13 and 2024-08-06) are marked Deprecated and you should not start new deployments on those. Embedding models are also available in Canada East under Standard: text-embedding-3-large, text-embedding-3-small, and text-embedding-ada-002.

Your observation on GPT-5 is correct. The gpt-5 family is not listed under Standard/Regional for any region, including Canada East. It is only available as Global Standard, Global Provisioned Managed, or Regional Provisioned Managed (PTU) in Canada East. There is currently no way to get gpt-5 as a pay-per-token Standard deployment with Canada-only processing. If you need gpt-5 with strict Canadian residency, Provisioned (PTU) in Canada East is the only path today.

Practical recommendation for your production design. Deploy gpt-4o (2024-11-20) or gpt-4.1-mini as Standard in Canada East now, since both are GA and not deprecated, but plan your migration ahead of their October 2026 retirement dates rather than treating them as long-term choices. Check the Model Retirement Schedule periodically since replacement models for the Standard/Regional category in Canada East tend to roll in gradually rather than all at once.

https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule

1. Verify current availability and quota before committing. Region availability and quota change over time, so confirm directly in the portal before finalizing your design. In the Foundry portal go to Operate > Quota, turn on Show all, and filter by Canada East to see live model and quota availability for your subscription.

https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability

1. If gpt-5 capability is a hard requirement, weigh Provisioned (PTU) in Canada East against your cost constraints, since that is the only way to combine gpt-5 with strict in-region processing today. There is no published timeline for when gpt-5 family models will be added to Standard/Regional in Canada, so this is not something you can plan around with certainty.

If this answer helps you kindly accept the answer which will help others who have similar questions.

Regards,

Sangeetha Kesavan