3 people reported

All time, since Sep 22.

Sub-agents use unselected models

“On 2026-09-23, three large usage entries appeared under gpt-5.6-sol-medium and consumed a large share of my Other Models budget (≈74%).”

Is this your bug? What gets counted gets fixed. Pile on.

Paste this into your agent. It files your report, with your error and version.

I hit this bug: rapsheet.dev/sheets/cursor/issues/2206/file
Follow it to file my report. Show me the text before you send.

No install. No account. Public and anonymous. The link asks your agent for one report, nothing else. What it sends

Quiet. No reports in the last two weeks, down from 2 a week at the late-September peak.

Sub-agents run on models the user did not select. Parent agent or background tasks launch expensive models like Opus or Grok, ignoring BYOK config and draining usage budget.

Is this your bug?

File my report
Where
Task sub-agents
Versions
3.21.16
Platforms
Windows
How bad
Breaks a normal workflow.

Not quite your bug?

  1. Subagents fall back to Auto11 people

Original reports3, newest first

  1. Bug: Single document edit request triggered massive Grok 4.6 Medium token consumption (834k tokens), ignoring BYOK model config#174391
  2. Unexpected gpt-5.6-sol-medium usage (~74% of Other Models budget) from Task sub-agents I did not select#172975
  3. 2 Claude Opus 5 sub agents drained 32M tokens... (~70% usage gone)#172673

Do you work on Cursor?

Claim its RAP Sheet to see what your users report and answer them in one place.