Describe the Bug
I have a cursor pro plus suscription. I also have a google One Account with Pro access to gemini. For a while now, Gemini has been failing a lot. It sometimes simply stop thinking mid process. The UI of how it shows the output is kind of broken (E.g. this is the beginning of an output “< / thought> Of course I will analyze the bug”). It does not follow instructions correctly, even when I provide extreme details. None of this issues happen when I use gemini in either its general app or AI studio. So not sure if this is a gemini issue or cursor issue.
I’ve been using claude in hopes it was going to get fixed but I ran out of credits and I need to continue working. What I now did is create a new API key in my google AI studio account and put it in cursor settings to use gemini until my period resets and can use claude again, pero the issue continues.
Steps to Reproduce
Use gemini 2.5 Pro in agent mode with user API key. Basically any request fails.
Expected Behavior
I would expect correct agentic behaviour without failing mid thinking, perform tasks and such.
Operating System
Mac
Current Cursor Version (Menu → About Cursor → Copy)
Version: 1.6.26 (Universal)
VSCode Version: 1.99.3
Commit: 6af2d906e8ca91654dd7c4224a73ef17900ad730
Date: 2025-09-16T17:12:31.697Z (23 hrs ago)
Electron: 34.5.8
Chromium: 132.0.6834.210
Node.js: 20.19.1
V8: 13.2.152.41-electron.0
OS: Darwin arm64 24.6.0
For AI issues: which model did you use?
Gemini 2.5 Pro
Does this stop you from using Cursor
No - Cursor works, but with this issue