Tibo
@thsottiaux
Codex team is aware of reports of GPT-5.5 performing worse for some users and investigating. We don't have anything conclusive yet and systems are healthy but we will share updates as we go.
167retweets5.5klikes
View on XOpenAI Investigates Reports of GPT-5.5 Performance Regressions in Codex
OpenAI is investigating reports of performance regressions for GPT-5.5 within its Codex coding platform. Tibo, the Codex engineering lead, confirmed the team is aware of users experiencing worse results. While the infrastructure remains healthy, there is no conclusive explanation for why the model is performing worse.
The investigation follows the recent OpenAI's GPT-5.5 launch, which introduced the model to the platform. While internal systems are healthy, the engineering team is prioritizing these reports to identify why the model is performing worse for some individuals. Consistent performance is essential for the platform's reliability as it handles complex coding workloads.
No immediate action is required from users, as the investigation is internal and systems are operational. You should monitor your coding sessions for unexpected errors or logic failures stemming from these performance reports. OpenAI will share updates as the investigation progresses through official developer channels and social updates from the engineering team.
This story has since progressed — see OpenAI Fixes GPT-5.5 Performance Issues and Resets Codex Usage Limits for the current state.
Still wondering? A few quick answers below.
OpenAI is currently investigating reports that GPT-5.5 is performing worse for some users within the Codex platform. While the engineering team has acknowledged these performance regressions, they do not yet have a conclusive explanation for the dip in quality. The team is actively looking into the issue to restore expected performance levels.
No, OpenAI has confirmed that its internal systems are currently healthy and operational. The investigation is focused on specific reports of GPT-5.5 performing worse than expected rather than a widespread service failure or technical outage. Users can still access the platform, though they may encounter inconsistent results during the ongoing investigation.
The Codex engineering team at OpenAI is actively investigating the reported performance regressions to identify a root cause. While the systems appear healthy, the team has committed to sharing updates as they gather more information. There is currently no timeline for a resolution, but the investigation is a priority for the engineering leads.
