I’m using Cursor on a personal account (not a team/workspace) and image generation is not available.
In Cursor Settings → Models, searching for image shows no results (0 image models).
When the agent tries to generate an image, it fails with something like:
“GenerateImage is not available for the current selected model.”
I expected image generation to be available (it’s mentioned in the Cursor changelog), but it looks like the image model is not exposed to my account/app.
Steps to Reproduce
Open Cursor
Go to Settings → Models
Search for image in the models list → no models appear
Ask the agent: “Generate an image …” → image generation tool is unavailable / fails
Expected Behavior
Image generation should be available (or at least image models should appear in Settings → Models).
On the first point, image generation models don’t show up in Settings → Models, that’s expected. Image generation runs through a separate internal tool that the agent calls automatically, so you don’t need to pick it manually in settings.
But if the GenerateImage tool isn’t triggering at all, that sounds like a feature access issue on your account. To dig in, I’ll need a Request ID:
Open a new chat in Agent mode
Ask it to generate a simple image, for example “Generate a blue circle icon”
When it fails, open the chat context menu top right > Copy Request ID
Send it here
That’ll let us check what’s blocking image generation on your account.
I checked your Request ID. The issue is that in Auto mode your request got routed to a different model that doesn’t support image generation.
Some models currently don’t support image generation:
gpt-5.1
gpt-5.2
gpt-5.3-codex
When you use Auto there’s a chance your request will end up on one of them.
Workaround: instead of Auto, manually pick a model that supports image generation, like Composer 1.5, before asking the agent to generate an image. That should fix it.
Nano Banana isn’t a model you can select. It’s an internal tool name for the image generation backend. When you use a model that supports image generation like Composer 1.5 and ask it to generate an image, it automatically uses that tool in the background.
So there’s nothing extra to set up. Just make sure you’re using a model that supports image generation, avoid gpt-5.1, gpt-5.2, and gpt-5.3-codex, then ask the agent to generate an image like usual.
