Subagents using wrong model

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Hello, when my agent is spawning Subagents , sometimes it uses models that it shouldn’t … In my settings I have it set for Subagents Models : Cursor Grok 4.5 … But sometimes my agent spawn subagents with GPT 5.6 Luna (which my API limit is already full and so if my budget wouldn’t be disabled it would be using it …) or even i’ve seen GPT 5.6 Sol too sometimes …

It can be annoying , more so when I have my On-Demand Usage enabled with a budget because it’s using it without me sometime even noticing …

Anyway, I hope this can be fix ASAP. Thanks and have a nice day :smiley:

Steps to Reproduce

Have your agent spawn subagents.

Expected Behavior

Use the right model set in the settings.

Operating System

Linux

Version Information

Version: 3.14.7
VS Code Extension API: 1.128.0
Commit: a758f2241ca99fecf380180b6cbdbbce0f1f42c0
Date: 2026-07-30T06:41:34.009Z
Layout: IDE
Build Type: Stable
Release Track: Default
Electron: 40.10.6
Chromium: 144.0.7559.236
Node.js: 24.15.0
V8: 14.4.258.32-electron.0
xterm.js: 6.1.0-beta.291
OS: Linux x64 7.1.2-1-cachyos-xefan

For AI issues: which model did you use?

I was using Cursor Grok 4.5 when it happened.

Does this stop you from using Cursor

No - Cursor works, but with this issue

Hey, thanks for the report. This is a known issue we’re tracking. The Subagents Models setting only sets the default, but when the parent agent spawns a subagent it can still pick a different model from the ones enabled in Settings > Models. That’s why it only happens sometimes.

Workaround for now: go to Settings > Models and disable the models you don’t want subagents to use GPT 5.6 Luna, GPT 5.6 Sol, etc. Disabled models disappear from the list available to subagents in the IDE, so the agent can’t choose them. One note: new models are enabled by default, so after updates it’s worth rechecking the list.

We also recently shipped a server-side change that more strongly nudges subagents to inherit your selected model, so this should happen less often. Full control over this is in progress too, but I can’t share an exact timeline. I’ll post here if there’s an update. Let me know if the workaround helped.

Thanks ! And yes , for the workaround that is exactly what I have done haha I thought about it right after creating the post haha !