Ultra: all Gemini models return provider 401

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

On an Ultra plan, every Gemini model fails immediately with ERROR_PROVIDER_ERROR and providerStatusCode 401. Other providers work.

Steps to Reproduce

  1. Google API Key off. OpenAI API Key off. OpenAI Base URL empty.
  2. New Agent chat.
  3. Select any Gemini model (Gemini 4 Argon, Gemini 3.8 Flash, or Gemini 3.7 Flash).
  4. Send “hi”.

Expected Behavior

Gemini replies through the included Ultra route.

Operating System

MacOS

Version Information

IDE:
Version: 3.27.22
VSCode Version: 1.128.0
Commit: b1aa50cb6db84fd3e8c808259cb172ea73425c10
Date: 2026-10-27T00:33:02.045Z
OS: macOS 15.5 (Build 24F74)
Quality: stable

The same 401 also happens on a second computer with the same Ultra account and the Google API Key left off.

For AI issues: which model did you use?

Gemini 4 Argon, Gemini 3.8 Flash, Gemini 3.7 Flash

For AI issues: add Request ID with privacy disabled

4756886c-6829-4913-98a0-634ce3f10a27
b9f6b284-19b5-4e98-a040-c265e34af278
bd3d8ce4-7b26-4c34-9877-8dc6feb8b083

Additional Information

A personal Google API key was tested only as a workaround. With the key on, requests got past the 401 and then failed with “User API Key Rate limit exceeded” on the free tier. Please restore Gemini on this account through the included Ultra route, not BYOK.

Email sent to [email protected] with subject “Ultra: all Gemini models return provider 401”.

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

Update: Gemini is working again on this Ultra account with the Google API Key turned off. Please treat this as resolved. I will reply here if the 401 returns.