alorbach
(Andre Lorbach)
1
Where does the bug appear (feature/product)?
Cursor IDE
Describe the Bug
Simple Tasks suck massive amounts of tokens in spending with “claude-fable-5-thinking-high” model.
Steps to Reproduce
select: claude-fable-5-thinking-high
let it do something simple like write hello world app. Then review in “Usage” how much tokens were burned.
Operating System
Windows 10/11
Version Information
Version: 3.7.27 (user setup)
VS Code Extension API: 1.105.1
Commit: e48ee6102a199492b0c9964699bf011886708ba0
Date: 2026-06-10T01:46:16.942Z
Layout: editor
Build Type: Stable
Release Track: Early Access
Electron: 39.8.1
Chromium: 142.0.7444.265
Node.js: 22.22.1
V8: 14.2.231.22-electron.0
xterm.js: 6.1.0-beta.256
OS: Windows_NT x64 10.0.26200
Does this stop you from using Cursor
Yes - Cursor is unusable
alorbach
(Andre Lorbach)
4
Just look
These are the IDE Session for these token usages:
I can see my API Spending quota melt in minutes
AndrewC
(Andrew)
6
It’s a massively thinking model… So it’s probably overthinking your simple tasks.
alorbach
(Andre Lorbach)
7
but why is the Context Usage so much less ?
AndrewC
(Andrew)
8
Also dont every use MAX mode!
choose your models for the right task.
use Fable for Plan mode, then use composer 2.5 for Agent mode.
You can also set the models Effort level …
alorbach
(Andre Lorbach)
9
I am just testing the new model, most work I am doing with composer-2.5.
These were my settings (The default model settings set by Cursor btw):
Does that mean max mode active all the time ?
Still it does not explain why so much tokens are being charged
AndrewC
(Andrew)
10
Id set the setting per modul yourself - Cursor might just set things that use lots of tokens
Colin
(Colin)
11
@alorbach A lot of what you’re seeing are cache tokens! Have a look at my explanation here.
