AI Incidents
Back to incident register
Source reviewedOperational incidentHigh

OpenAI outage reaches a 35% ChatGPT error rate

A routine operating-system update disconnected numerous GPU nodes from the network. ChatGPT errors reached roughly 35%, while API errors reached roughly 25%.

First observed
Jun 10, 2026, 8:36:00 AM
Disclosed
Jun 10, 2026, 8:36:00 AM
Status
Resolved
Confidence
99%
Organization
OpenAI
Last reviewed
Sep 1, 2026

Observed

Facts supported by sources

  • ChatGPT errors peaked at approximately 35%, while API errors peaked at approximately 25%.
  • According to the postmortem, API availability fell to 75%.
  • The UTC times were converted from PDT.

Assessment

Interpretation, not observation

No additional interpretation recorded.

Impact

Impact

Broad global outage with a high error rate; full recovery took approximately 15 hours and 24 minutes.

Response

Response

OpenAI reimaged GPU nodes, stopped automatic updates, and added configuration audits and disaster-recovery exercises.

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