Back to wire
AI·Article·Resolved

ChatGPT Work recovers after an 81-minute partial outage

OpenAI says ChatGPT Work has fully recovered after a partial outage in which some users could not start or run Work tasks. The incident ran from 19:03 to 20:24 UTC on 17 September; OpenAI identified and mitigated the failure but has not published a root cause.

Published 17 Sept 2026, 21:03 · Updated 17 Sept 2026, 22:24

Work tasks failed for some users

OpenAI opened the incident at 19:03 UTC on 17 September after some users began encountering failures when starting or running tasks in ChatGPT Work mode. The status record classifies the event as a partial outage and identifies ChatGPT Work as the affected component. The incident record does not mark the API, ordinary ChatGPT conversations, Codex or other ChatGPT components as affected.

The service recovered at 20:24 UTC

OpenAI moved from investigation to an identified issue at 19:57 UTC and said it was implementing a mitigation. At 20:24 UTC, the company marked the incident Resolved and said all impacted services had fully recovered. The recorded impact window was about 81 minutes from the first incident timestamp to recovery.

Root cause remains undisclosed

OpenAI has not published a technical cause, a quantified share of affected users or a post-incident explanation. The resolved status establishes that the service degradation ended, but it does not support broader conclusions about why the failure occurred or whether it was related to other ChatGPT infrastructure.

Source trail

1 source · 1 primary