Codex (Chatgpt) Desktop Runs via xWayland at 3.25 hz

Open
#39,462 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
48/100
Issue type
Bug
Clarity
Mostly clear
Activity status
Active
Tech stack
electron, linux
Domain
desktop, operating-systems

Research direction

Start by examining how the Linux desktop launcher selects Electron's X11 or Wayland mode for GNOME Wayland multi-monitor sessions. Reproduce the lag with the native .deb package and compare it with chatgpt --ozone-platform=wayland --enable-features=WaylandWindowDecorations. Done means the launcher no longer forces X11 or disables GPU compositing in the affected configuration, without the severe presentation lag.

Written by the indexing model from the issue text.

Description

app bug Linux performance
What version of the Codex App are you using (From “About Codex” dialog)?

26.814.41957

What subscription do you have?

Pro

What platform is your computer?

Linux 7.0.0-30-generic x86_64 unknown

What issue are you seeing?

The Linux ChatGPT/Codex desktop app has severe UI lag under a GNOME Wayland multi-monitor session when it launches through XWayland. All UI operations are affected, including scrolling, typing, window interactions, and Codex thought-chain/progress displays.

Environment:

  • GNOME Wayland session
  • Primary: AORUS, 3840×2160 at 240 Hz, 150% scaling
  • Secondary: ASUS, 6016×3384 at approximately 60 Hz, 266.67% scaling
  • Current native .deb package: 26.814.41957
  • Electron 42.x builds are affected
  • The 2026-05-14 build, using Electron 41.2.0, does not exhibit the problem

The affected launcher detects a GNOME Wayland multi-monitor session and forces --ozone-platform=x11, along with disabled GPU compositing. Under XWayland, the app was associated with a 7680×4320 virtual output reported at only approximately 3.25 Hz. This appears to explain the extremely slow UI presentation.

Launching the same package with native Wayland completely resolves the lag:

chatgpt --ozone-platform=wayland --enable-features=WaylandWindowDecorations

After doing this, the main process, GPU process, and renderer processes all use --ozone-platform=wayland. The app no longer appears in the X11 client list, uses /dev/dri/renderD128, and is presented on the primary display at 240 Hz.

I made the workaround persistent using a user-level chatgpt.desktop override. Ideally, the launcher should not automatically force X11 or disable GPU compositing on GNOME Wayland multi-monitor configurations, particularly when the resulting XWayland output reports an invalid or extremely low refresh rate.

What steps can reproduce the bug?

As described above. xWayland is suspected to be the culprit.

What is the expected behavior?

No response

Additional information

No response

Dominant language
Rust
Stars
125k
Forks
19.5k
Avg merge
1m
Merged PRs (30d)
1k

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from openai/codex

All issues in openai/codex

Similar issues

More Rust issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.