gpt-5.6-sol medium on Codex: mixed languages
A user encountered an issue while developing a project with Codex using "gpt-5.6-sol" on medium reasoning effort. After instructing Codex to test and fix bugs, and then to pause at a checkpoint due to approaching a 5-hour limit, the remaining time limits rapidly decreased to zero. Simultaneously, Codex produced an answer containing words from different languages, leading the user to question its reliability for coding tasks.
This report details an unusual failure mode for "gpt-5.6-sol" on Codex, where it produced mixed-language output unlike typical code generation errors.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Sep 13, 2026, 11:48 UTC
IngestedOffset at this time: UTC+0Sep 13, 2026, 15:00 UTC
- Published
- Sep 13, 2026, 11:48
- Ingested
- Sep 13, 2026, 15:00
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
I'm developing a small project with Codex using gpt-5.6-sol on medium reasoning effort. Specifically, I instructed Codex to test the program and iterate until the bugs found were fixed. Since I was approaching the 5h limit (about 16% left), I told Codex to pause to a safe checkpoint. After that, the limits spiraled down to 5%, 2% and then suddenly zero, all while Codex produced an answer with words from different languages... has this happened to anyone before? My worry now is that I can't really trust it to code since how can I rule out that that's not going to happen while writing code?
https://preview.redd.it/ikqyrhlqz9ph1.png?width=1702&format=png&auto=webp&s=f8d3f4e06025c8c77c64d2473337720d6a9fd3be