Overview
- Thousands of users worldwide reported errors, loading failures, missing chat history and stalled prompts during a surge of complaints early Saturday.
- Downdetector and similar trackers recorded complaint totals rising into the thousands with regional clusters in the United States, Europe, India, Japan and Australia.
- Developers said failures of the OpenAI API and Codex left third‑party apps and business workflows unable to function or experiencing high latency.
- OpenAI acknowledged elevated error rates on its official status page, said mitigations were in place and engineers were monitoring recovery, but it has not identified a cause or given a timeline for full restoration.
- The outage follows a separate July 23 incident that took roughly 24 hours to resolve, sharpening concerns about reliability as more companies build products on a small set of AI platforms.