Start any chat with Grok 4.5 High and see that it burns tokens like h*ll in version
Version: 3.20.17 (system setup)
VS Code Extension API: 1.128.0
Commit: 0c32194e3fb5ffaced9fb36430b860ec301e1fc0
Date: 2026-09-12T03:16:10.634Z
Layout: IDE
Build Type: Stable
Release Track: Default
Electron: 42.10.0
Chromium: 148.0.7778.280
Node.js: 24.18.1
V8: 14.8.178.38-electron.0
xterm.js: 6.1.0-beta.291
OS: Windows_NT x64 10.0.26200
See that it burns millions of extra tokens like crazy and hallucinates also most of the times.
This behaviour is new since a couple of days now. Please fix ASAP
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!
We checked Grok 4.5 High usage across versions and 3.20.17 isn’t using more tokens than earlier builds. To look at your specific case, grab a request ID from a heavy chat (click the … at the top right of the chat, then Copy Request ID) and paste it here. If your Cursor login email differs from your forum one, tell me which to check.
Quick tip meanwhile: in Agent mode every tool call re-sends the whole conversation, so long multi-step chats pile up millions of tokens (mostly cheap cached reads). Starting a fresh chat per task cuts that a lot.
And Grok 4.5 is still selectable in Settings → Models, so it isn’t being swapped to 4.6 on you.