AI Incidents
Back to incident register
Source reviewedOperational incidentMedium

OpenAI reported elevated errors across ChatGPT and Codex for more than two hours

OpenAI documented elevated errors affecting ChatGPT and Codex on September 3. The incident lasted a little over two hours, and OpenAI warned that some Codex remote-control users might need to pair their mobile devices again afterward.

First observed
Sep 3, 2026, 4:43:00 PM
Disclosed
Sep 3, 2026, 4:58:23 PM
Status
Resolved
Confidence
99%
Organization
OpenAI
Last reviewed
Sep 4, 2026

Observed

Facts supported by sources

  • OpenAI reported elevated errors across ChatGPT and Codex from 14:43 UTC until resolution at 16:55 UTC on September 3.
  • OpenAI stated that some Codex remote-control users might need to pair their mobile device again following the incident.

Assessment

Interpretation, not observation

  • The public status record does not provide a root cause or affected-user count.
  • The incident overlapped with outages at other AI providers, but no shared cause is established by the source.

Impact

Impact

Failed ChatGPT and Codex requests and possible re-pairing work for affected Codex remote-control users.

Response

Response

OpenAI applied a mitigation, monitored recovery and then reported the issue causing elevated errors resolved.

Methodology: Facts, interpretations, and uncertainty are kept separate. Confidence describes the strength of the evidence, not a probability estimate.