An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
Hi Deep Mehta,
What you’re seeing is expected behavior with gpt‑5.2‑chat in Azure AI Foundry right now, and it’s a bit confusing because of how the portal labels capabilities.
Although the model catalog and portal description mention “agent / tool” scenarios, gpt‑5.2‑chat does not currently support tools when used inside Foundry agents. When you attach the model to an agent, Foundry correctly warns you and removes any configured tools. That warning is accurate and reflects the current runtime behavior, not a misconfiguration on your side.
A few important clarifications:
• The model itself supports tool‑style reasoning in some contexts, but
• The Foundry Agents execution path does not yet enable tool calling for gpt‑5.2‑chat
• This is why tools work with models like GPT‑4.1 or earlier GPT‑5 variants, but are disabled for gpt‑5.2‑chat in agents
• The portal text is ahead of the current Foundry agent runtime support
So there’s nothing broken in your setup. This is a feature gap, not a permissions or configuration issue.
Current practical options:
• If your agent requires tools or function calling, continue using GPT‑4.1 or another tool‑enabled model
• Use gpt‑5.2‑chat for chat‑only or multimodal scenarios where tools are not required
• If you’re calling the Responses API directly (outside Foundry Agents), you may see different capabilities than what Agents currently support
At the moment, there’s no published timeline in Microsoft Learn documentation for when gpt‑5.2‑chat tool support will be enabled in Foundry agents.
I Hope this helps. Do let me know if you have any further queries.
Thankyou!