Model Router with GPT-5.6 family

Hello @Nhan Hoang

Thank you for your Reaching out to Microsft Q&A.

Based on the current Azure AI Foundry Model Router documentation, the GPT-5 models that are currently supported as Model Router pool options are:

• gpt-5-nano (version 2025-08-07)
• gpt-5-mini (version 2025-08-07)
• gpt-5 (version 2025-08-07)
• gpt-5-chat (version 2025-08-07)

These models are listed as supported under the Model Router version 2025-11-18 documentation.

Regarding your specific question about the GPT-5.6 family, the current documentation does not list any GPT-5.6 models (including Terra or Sol variants) as supported Model Router pool options. As a result, GPT-5.6 models cannot currently be selected as routing targets through Model Router.

At this time, Microsoft has not published any roadmap or announced an ETA for when (or if) the GPT-5.6 family will be added to the Model Router supported model pool. Therefore, we are unable to confirm a timeline for this capability.

If your application requires GPT-5.6 today, the recommended approach is to deploy the GPT-5.6 model directly in Azure AI Foundry and invoke that deployment explicitly, rather than using Model Router.

We recommend monitoring the following resources for future announcements and updates:

• Azure AI Foundry Model Router documentation – for the latest list of supported routing models.
• https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/model-router?tabs=foundry-responses

If support for the GPT-5.6 family in Model Router is important for your workload, we also encourage submitting a feature request through the Azure Feedback portal. Customer feedback is regularly reviewed by the product team and helps inform future product planning.

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!

GPT-5.6 is not currently an underlying option for Model Router. Microsoft’s published supported-model table for the active 2025-11-18 router version lists GPT-5.5 and earlier supported GPT families, but not GPT-5.6, and Microsoft has not published an inclusion date. Availability of GPT-5.6 as a direct Foundry deployment does not automatically make it eligible for routing because the router’s pool is separately validated and released.

You do not need to wait for a new router version number: Microsoft states that 2025-11-18 is maintained in place as new models are added. Check the supported-model table and Model Router “What’s new” page for the authoritative change. If you use a custom model subset, add GPT-5.6 after it appears; new models are not inserted into existing subsets automatically. Until then, deploy GPT-5.6 directly where available. Microsoft Support or your account team can discuss roadmap information, but there is no documented date to guarantee.