OpenAI confirmed a temporary service disruption affecting ChatGPT Work Mode on September 3, 2026, resulting in elevated error rates across parts of the ChatGPT service.
According to OpenAI's official status page, the incident was registered under the title “ChatGPT Work Mode High Error Rates.” In its initial update, OpenAI stated that ChatGPT Work Mode was experiencing high error rates and that its technical team was investigating the issue.
The official incident timeline shows that the Investigating status was posted at 12:04 AM on September 3. Only a few minutes later, at 12:10 AM, OpenAI marked the incident as Resolved and reported that all impacted services had fully recovered.
OpenAI classified the event as a “Partial outage.” The status page also identified ChatGPT and multiple related components as affected during the incident.
Although users may encounter different access or application errors during service disruptions, OpenAI's official incident report did not identify a specific HTTP error code, including HTTP 404, as part of this event. As a result, individual 404 responses cannot be conclusively attributed to this incident without additional technical information.
The incident also highlights the infrastructure complexity behind modern AI platforms. Services such as ChatGPT depend not only on large language models but also on cloud infrastructure, backend services, authentication systems, networking, routing, data storage and APIs. A temporary issue within any of these layers can result in elevated error rates or reduced service availability.
The short interval between OpenAI's initial investigation notice and the recovery update indicates that the disruption was identified and resolved within a limited period. Following recovery, OpenAI confirmed that all affected services had returned to normal operation.
The incident is currently listed as Resolved on the official OpenAI Status platform, with the impacted services reported as fully recovered.