Cursor began printing Gemini 3 Pro's thoughts as plain text in the chat

[Igor_Markin]

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Cursor began printing Gemini 3 Pro’s thoughts as plain text in the chat

Steps to Reproduce

Working with Gemini 3 Pro for a long time

Screenshots / Screen Recordings

Operating System

MacOS

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.0.77 (Universal)
VSCode Version: 1.99.3
Commit: ba90f2f88e4911312761abab9492c42442117cf0
Date: 2025-11-13T23:10:43.113Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.1.0

For AI issues: which model did you use?

Gemini 3 Pro

For AI issues: add Request ID with privacy disabled

0d9abd9c-e34a-4c4b-b8c7-3eccaf5f0127

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

[deanrie]

Hey, thanks for the report. This is a known issue with Gemini 3 Pro: internal system instructions are “leaking” into the chat as plain text that looks like thoughts/reasoning. The team is already tracking it.

Since Gemini 3.0 Pro was launched only yesterday, the team is still working through several integration issues. The good news is that the model should still function correctly: the leaked text is just visual noise in the response and doesn’t affect actual code generation.

[liquefy]

As an addition, in case nobody has reported it yet, gemini in plan mode first makes changes to the codebase and only then makes the plan, not sure why, happens a lot.