It’s usually a version mismatch inside the AI Agent node, not n8n itself.
Even if your n8n core is updated, the Agent node doesn’t always refresh unless the instance is rebuilt cleanly. A quick fresh pull / redeploy without cache normally updates the node and removes that “model not supported” error.
If you tell me how you’re hosting n8n, I can point you to the exact fix so you get it sorted fast.
Thank you, I solved it, it’s just because “Use Responses API” was turned on
I’m having the same issue. Where do I find the Use Responses API toggle? Thanks in advance