Grok 4.5 & Composer 2.5 image generation fails: GenerateImage blocked as "unintended" despite explicit user request

Support Bug Reports
, ,
1

Where does the bug appear (feature/product)?

  • Editor, Tab & Chat (autocomplete, Composer, in-editor agent)

Describe the Bug

When I explicitly ask the Agent to generate an image (e.g. “Generate a picture of a cute kitten”), the Agent invokes GenerateImage but it always fails. Image generation used to work with Composer 2.5; it stopped working recently.

The Agent-side error is:

Failed to generate image, error: Image generation was not performed — the tool call appears to be unintended. Do not retry or call GenerateImage again unless the user explicitly asks for an image.

The user intent is unambiguous, yet the call is rejected as “unintended.”

Steps to Reproduce

  1. Open Cursor and open a project folder (not Quick Chat).
  2. Start a new chat (Ctrl+N).
  3. Manually select Composer 2.5 (Auto / Premium model router disabled).
  4. Prompt: “Generate a picture of a cute kitten.”
  5. Agent calls GenerateImage → fails with the error above.

Also reproducible with Grok 4.5 (same behavior).

Expected Behavior

GenerateImage should run successfully and return an inline image preview (saved to the project assets folder by default).

Screenshots / Screen Recordings

N/A — no visible error dialog in the client UI; failure only appears in the Agent tool result.

Operating System

  • Windows 10/11 (Windows 10, build 10.0.26200)

Version Information

Version: 3.15.6 (user setup)

VS Code Extension API: 1.128.0

Commit: a1f686545fd0ce8917bbd2449f733551a9bce420

Date: 2026-08-06T01:41:03.876Z

Layout: IDE

Build Type: Stable

Release Track: Default

Electron: 40.10.3

Chromium: 144.0.7559.236

Node.js: 24.15.0

V8: 14.4.258.32-electron.0

xterm.js: 6.1.0-beta.291

OS: Windows_NT x64 10.0.26200

For AI issues: which model did you use?

Composer 2.5 (manually selected). Also tested Grok 4.5 — same failure.

For AI issues: add Request ID with privacy disabled

Request ID: Not available — the chat menu does not show “Copy Request ID” for these failed attempts, and there is no client-side error dialog.

If needed, I can disable Privacy Mode (Settings → General) and retry to obtain a Request ID. Please advise if that is required.

Additional Information

Troubleshooting already tried:

  • Brand-new chat
  • Sign out and sign back in
  • Manual model selection (Composer 2.5, not Auto)
  • Project folder open

No VPN / custom API base URL / BYOK configured: Yes

This appears to be a server-side routing or GenerateImage guard issue rather than a local configuration problem.

Does this stop you from using Cursor?

  • No - Cursor works, but with this issue
2

Hi there!

We detected that this may be a bug report, so we’ve moved your post to the Bug Reports category.

To help us investigate and fix this faster, could you edit your original post to include the details from the template below?

Bug Report Template - Click to expand

Where does the bug appear (feature/product)?

  • Editor, Tab & Chat (autocomplete, Composer, in-editor agent)
  • Terminal & commands
  • Models, pricing & API keys (availability, Auto/Max, BYOK/Bedrock)
  • MCP & tools
  • Cloud Agents & Automations (cursor.com/agents, scheduled/event)
  • BugBot & Code Review
  • Cursor CLI
  • Cursor Mobile
  • Remote (SSH / Dev Containers / WSL)
  • Account, billing & login
  • Something else…

Describe the Bug
A clear and concise description of what the bug is.

Steps to Reproduce
How can you reproduce this bug? We have a much better chance at fixing issues if we can reproduce them!

  1. …
  2. …
  3. …

Expected Behavior
What is meant to happen here that isn’t working correctly?

Screenshots / Screen Recordings
If applicable, attach images or videos (.jpg, .png, .gif, .mp4, .mov)

Operating System

  • Windows 10/11
  • MacOS
  • Linux

Version Information

  • For Cursor IDE: Menu → About Cursor → Copy
  • For Cursor CLI: Run agent about in your terminal
IDE:
Version: 2.xx.x
VSCode Version: 1.105.1
Commit: ......

CLI:
CLI Version 2026.01.17-d239e66

For AI issues: which model did you use?
Model name (e.g., Sonnet 4, Tab…)

For AI issues: add Request ID with privacy disabled
Request ID: f9a7046a-279b-47e5-ab48-6e8dc12daba1
For Background Agent issues, also post the ID: bc-…

Additional Information
Add any other context about the problem here.

Does this stop you from using Cursor?

  • Yes - Cursor is unusable
  • Sometimes - I can sometimes use Cursor
  • No - Cursor works, but with this issue

The more details you provide, the easier it is for us to reproduce and fix the issue. Thanks!

7

Hey there! Thanks for the detailed report.

What you’re seeing isn’t intended behavior, and it’s not anything in your setup. Quick unblock: give a more detailed image description in your prompt, e.g. Generate a picture of a cute kitten sitting in a basket, photorealistic, soft lighting. A fuller description reliably goes through - very short ones are what trip it up. It tends to show up more with Grok 4.5 and Composer 2.5 since they often phrase the image request more tersely.

We’ve let the team know and this is an issue we’re tracking. I’ll post here when there’s an update.

8

Thank you for your prompt reply
Today, I noticed that the cursor had been updated to version 3.15.19, and after the update, I found that the issue with image generation seemed to have improved. Even with the simplest prompts, such as “generate an image of a puppy”, it could generate the desired image.
But sometimes there are still problems, and here are a few phenomena I have observed:

  1. Using the GLM5.2 model, if the image generation tool is unavailable, the task will be stopped directly with a prompt
  2. If using the grok or composer model, the task will not stop and will continuously display “planning next moves”
  3. When I used to create page prototypes, I could have the cursor generate five images simultaneously for my confirmation. Now, when I use the same method, the first two images often generate, but the rest get stuck