Didn't follow directions, circular logic, filled up context with docs I didn't tell it to read.
Resolved 💬 3 comments Opened Dec 12, 2025 by kasselvania Closed Dec 12, 2025
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
codex 0.71.0
What subscription do you have?
API
Which model were you using?
GPT 5.2
What platform is your computer?
MacOS
What issue are you seeing?
Used the highest quality model. Asked it to read two guidance documents and edit a script to migrate from a v1 standard to a v2 standard. Filled its context twice reading many many other documents. Each time, it did minimal code edits, but didn't record its progress in the TODO, so after each context compact, it began from the start. Ate $10 of API costs with literally no good value.
What steps can reproduce the bug?
Uploaded thread: 019b1399-cc43-7850-805d-8abcb675d1c1
What is the expected behavior?
To read what I told it to and just edit the code.
Additional information
_No response_
3 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Just ran cursor using opus 4.5 and it found multiple mistakes made by the codex (including missed functions it should have audited) within moments.
Thanks for the bug report. We'll share the feedback with the team responsible for training our codex models. We're tracking this feedback and similar reports so we can make future models more robust and efficient. And thanks for using the
/feedbackcommand. That additional information is useful.