Seemingly degraded/slowed Codex performance in large(r) projects
Resolved 💬 4 comments Opened Nov 20, 2025 by moostoet Closed Mar 29, 2026
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
codex-cli 0.60.1
What subscription do you have?
PRO
Which model were you using?
gpt-5.1-codex-max
What platform is your computer?
Windows 11, using WSL Ubuntu 24.04
What issue are you seeing?
There seems to be a degraded performance across specific models during my sessions compared to a few days ago. Not much else to say on this other than that it took about 5 minutes for it to process a 2-line change in a file it already worked on with clear instructions.
What steps can reproduce the bug?
Uploaded thread: 019aa082-a39b-7b61-92b6-45cc7f8256d7
What is the expected behavior?
I expect there to be at least a better degree of performance for smaller tasks such as these. (Slightly related might be the ghost_commit problem, though I do not believe that it is enabled by default on projects).
Additional information
_No response_
4 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
gpt-5.1-codex-maxgot released just 12 hours ago - what are you comparing this against?Are you using
gpt-5.1-codex-maxwith extra-high reasoning? just came out last night - and it's advertised as being deliberately _slower_ (i.e. takes longer to reason)Confirmed. GPT-5.1-Codex-Max shows terrible performances even on Extra high. Furthermore, I found out that it is automatically switching to model such as GPT 4o even if in the selections says otherwise. GPT 5 Codex was mostly ok for me except the annoying “Stream disconnected before completion error”.I am using it in VS code.
This bug report hasn't received any upvotes or additional posts in three months. Closing.