Codex Mac app high memory usage regression on macOS after update (26.422.62136)
What version of the Codex App are you using (From “About Codex” dialog)?
Codex on Mac 26.422.62136
What subscription do you have?
ChatGPT Plus
What platform is your computer?
macOS Tahoe 26.4.1 (25E253)
What issue are you seeing?
I am seeing a major memory usage regression in the latest Codex Mac app version.
On the latest version, Codex memory usage grows to around 6 GB RAM or more during a normal coding session. This causes the app and the whole system to become noticeably slow and laggy.
I tested the same machine and the same projects with an older Codex Mac build from February 12. On that older version, Codex uses around 1.5 GB RAM and everything works smoothly. Projects open almost instantly and the app remains responsive.
After returning to the latest version, the high memory usage and lag come back.
This looks like a regression introduced in one of the recent Codex Mac app updates.
What steps can reproduce the bug?
- Open the latest Codex desktop app on macOS.
- Open a normal coding project.
- Start a normal coding session.
- Leave Codex open during regular work.
- Observe that memory usage grows to around 6 GB RAM or more and the app/system becomes slow.
- Compare with an older Codex Mac build from February 12 on the same machine and projects — memory usage stays around 1.5 GB RAM and the app remains responsive.
What is the expected behavior?
I expect Codex memory usage to stay at a reasonable and stable level during normal use.
The app should remain responsive on the same machine and projects that worked smoothly in previous versions.
Additional information
I tested an older Codex Mac build from February 12: 260212.1136
With the older build, Codex uses around 1.5 GB RAM and works smoothly on the same machine and projects.
With the latest build, Codex can use around 6 GB RAM or more, causing significant lag.
This strongly suggests that the issue is related to a recent Codex Mac app update rather than my local setup.
This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗