Are GPT-4o-mini and other models retiring from Azure Foundry as well on March 31st, 2026?

Question

OpenAI is retiring the GPT-4o model from ChatGPT, along with several other older models (GPT-4.1, GPT-4.1 mini, and o4-mini).

• Retirement Date: February 13, 2026
• Also, Got a notification that, 40-mini retires on March31st, 2026. But does it mean it gets retried from Azure Foundry as well? Please answer, I need to switch it before my prod goes down.

Answer 1

Hi Patel, Hardip

Azure AI Foundry deployment is using GPT‑4o or GPT‑4o‑mini (or “4o‑mini / 40‑mini”) in Standard deployments, it will stop working once the retirement date for that model/version is reached. The ChatGPT retirement date (Feb 13, 2026) and the Azure Foundry retirement date (Mar 31, 2026) are different timelines, and Azure Foundry is what matters for your production workload [learn.microsoft.com],

ChatGPT retirement ≠ Azure Foundry retirement:

The OpenAI announcement dated Feb 13, 2026 applies only to ChatGPT (UI experience). OpenAI explicitly states that there are no immediate API changes tied to that ChatGPT retirement [openai.com]

Azure AI Foundry has its own retirement schedule:

Azure OpenAI models in Microsoft Foundry follow Azure‑specific lifecycle rules. Once a model is retired in Foundry, all requests return errors [learn.microsoft.com]

Microsoft has confirmed that the following Standard deployments are retired on March 31, 2026:

• gpt‑4o (2024‑05‑13)
• gpt‑4o (2024‑08‑06)
• gpt‑4o‑mini / 4o‑mini (commonly shown as “40‑mini” in notifications)

If your deployment is pinned to one of these versions, your prod will break on that date [learn.microsoft.com]

you received the “40‑mini retires March 31” notification:

That notification is coming from Azure Resource Health / Service Health, not OpenAI. Azure sends these per‑deployment and per‑SKU once a model version reaches its retirement window [learn.microsoft.com]

Deployment type matters (very important)

The March 31, 2026 retirement applies to:

• Standard deployments

Later retirement (October 1, 2026) applies to:

• Provisioned
• Global Standard
• Data Zone Standard

If you are running Standard (most common for PoC and early prod), you must migrate before March 31

you should do immediately to avoid prod outage:

You should not wait.

Safe actions:

• Switch deployment to gpt‑5.1 (2025‑11‑13) or another currently GA model
• Ensure your app references the deployment name, not a hard‑coded model version
• Set upgrade policy to: “Stay on this version” only if you’ve already switched or “Upgrade when current version expires” if validated Failure to do this results in hard 500/404 errors on March 31 [learn.microsoft.com]

Relevant Documentation:

• Azure OpenAI Service Model Version Deprecation
• Working with Azure OpenAI Models

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!

Answer 2

Yes, the gpt-4o-mini model is set to retire from Azure Foundry as well. According to the information available, the gpt-4o-mini model has a retirement date of March 31, 2026. Therefore, you will need to switch to a different model before this date to avoid any disruption in your production environment.

---

References:

• Azure OpenAI in Microsoft Foundry model deprecations and retirements
• Retired Azure OpenAI models in Microsoft Foundry
• Azure OpenAI in Azure AI Foundry Models retired models
• Azure OpenAI in Microsoft Foundry model deprecations and retirements