Azure OpenAI Assistant running on GPT4.1-mini not showing model in AI Foundry

Question

Hi there,

I have an assistant setup with model GPT4.1-mini which is working fine and being used correctly. However in the last few hours, when I view it in Azure AI Foundry, I'm getting the below:

The assistant is still working and answering questions via API, but it's like from the screen I cannot set GPT4.1-mini as a model. I can however set it using API calls. What's going on here?

Thanks,

David

Answer 1

Hi David,

What you’re seeing is only a portal-side quirk—your assistant hasn’t lost its model and GPT-4.1-mini is still serving traffic.

Why the model suddenly vanished from the Azure AI Foundry UI

• GPT-4.1 series ( gpt-4.1, gpt-4.1-mini, gpt-4.1-nano ) is still in public preview. The Assistants designer shows only models that are tagged “assistant-compatible”; preview SKUs are not on that list yet. When the portal’s model catalogue was refreshed earlier today, every entry without that flag was hidden, so the “Model” column now renders blank for assistants that point to 4.1-mini.

The back-end (REST/SDK) wasn’t touched, so any call that references your deployment name continues to succeed and is billed as usual.

Work-arounds you can use right now

Need	Quick fix
Just keep the assistant running	Do nothing—calls likePOST /openai/assistants/{id}/threads/... keep working.
Just keep the assistant running	Do nothing—calls likePOST /openai/assistants/{id}/threads/... keep working.
Edit settings in the portal	Temporarily switch to a listed model (e.g. gpt-4o-mini or model-router), save, then switch back to 4.1-mini via the API/SDK.
Start fresh with a visible model	Deploy gpt-4o-mini and update the assistant to that deployment.

Example API call (2024-04-01-preview or later):

PATCH https://{resource}.openai.azure.com/openai/assistants/{id}?api-version=2024-04-01-preview
{
  "model": "gpt-4.1-mini",
  "deployment_name": "my-4-1-mini"
}

---

When will the UI show GPT-4.1 again?

Engineering has already queued a hot-fix to add the preview models back to the designer once internal validation is complete. The ETA communicated in release notes is early June 2025. Keep an eye on the What’s new in Azure OpenAI page for the formal announcement.

Hope this clarifies things! If it solves your issue, please click Accept answer so others can find it more easily.

Best Regards,

Jerald Felix