Bug Report: Chinese Input Character Loss in Codex CLI
Resolved 💬 2 comments Opened Aug 28, 2025 by penkzhou Closed Oct 18, 2025
💡 Likely answer: A maintainer (joshka-oai, contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
0.25.0
Which model were you using?
gpt-5
What platform is your computer?
Darwin 24.5.0 arm64 arm
What steps can reproduce the bug?
- Open Codex CLI.
- Input 4 Chinese characters in one keystroke (e.g., “测试输入”).
- Observe that only 2 characters are displayed.
- Continue typing more characters.
- Observe that the missing characters appear retroactively.
What is the expected behavior?
All input Chinese characters should appear immediately without being lost or delayed.
What do you see instead?
Some characters are not shown at first (e.g., only half are displayed), but they are restored once additional typing occurs.
Additional information
Codex CLI version: 0.25.0
OS: macOS 15.5
Terminal: iTerm2
Input method: macOS built-in Chinese Pinyin IME
2 Comments
Hey, it looks like this is probably already fixed in the lates codex cli (0.47) on macOS. Can you please confirm?
I selected Pinyin - Simplified and entered ce shi shu ru (via https://chatgpt.com/share/e/68f2e93b-2808-8009-80c8-0314343d5f5b) and hit space to insert and saw 4 characters appear. Backspace works correctly and future characters continue to work.
I'm working on a remote machine using ssh, and the system is Ubuntu 22.04.5 LTS. Chinese characters disappear after backspacing.
https://github.com/openai/codex/issues/5318
https://github.com/user-attachments/assets/b35880d9-725c-4ae0-92d1-c17a2e2d6d2d