OpenAI Outage History

The OpenAI /v1/videos endpoint experienced elevated errors, impacting Sora / Video generation. Users reported that Sora 1 could not be accessed during the incident. The outage began at 3:41 AM UTC and remained in a degraded state until 4:51 AM UTC when it was restored, lasting approximately 1 hour 10 minutes.

OpenAI issued 2 status updates during the incident as mitigation was applied. The issue affected users across multiple locations with video generation features impacted, and service returned to normal after the mitigation was implemented.

OpenAI's Sora video generation endpoint (/v1/videos) experienced degraded performance and high error rates beginning around 2:04 AM UTC as heavy load impacted video creation. Users across North America reported being unable to generate videos, with processing delays and errors. StatusGator detected the issue at Oct 24, 2025 2:04 AM UTC, about 37 minutes before OpenAI acknowledged the problem on their status page.

The incident lasted from 2:04 AM UTC to 4:51 AM UTC (roughly 2 hours and 47 minutes). OpenAI issued 2 status updates during the event: first at 3:41 AM UTC indicating a degraded condition for the Sora / Video generation component, and a follow-up at 4:51 AM UTC signaling recovery as service returned to normal. Reports showed a North America pattern with widespread heavy load and video processing failures, though the impact did not constitute a full outage.

OpenAI experienced a brief degraded performance affecting the ChatGPT / Connectors component due to issues with the Company Knowledge feature. The incident began at 7:31 PM UTC when the status changed from up to a degraded state, with a message indicating issues with Company Knowledge and that investigators were looking into it. The impact appears limited to the ChatGPT / Connectors functionality.

OpenAI reported elevated errors affecting ChatGPT / Conversations, impacting user interactions across the ChatGPT conversations component. The outage began at 3:11 PM UTC and was resolved by 4:11 PM UTC, with users experiencing difficulties generating or receiving responses and frequent error messages. A surge of user reports in the mid-UTC afternoon indicates broad impact on conversations.

OpenAI issued two official updates during the incident. The first noted elevated errors and that mitigation was underway; the second confirmed restoration of services. The disruption appeared to affect users across multiple regions, suggesting broad reach rather than a single-region issue. Functionality resumed after the resolution.

OpenAI ChatGPT conversations experienced elevated errors and degraded performance, with users reporting failures to receive responses and error messages across multiple regions. StatusGator detected reports starting at Oct 23, 2025 2:44 PM UTC, about 27 minutes before OpenAI acknowledged the issue on their status page. The outage lasted from Oct 23, 2025 2:44 PM UTC to Oct 23, 2025 4:11 PM UTC (roughly 1 hour and 27 minutes). Reports indicate widespread impact across users in multiple regions rather than a single geography.

OpenAI's GPT-4.1-Nano APIs, including the Chat Completions and Responses endpoints, experienced elevated 503 errors that disrupted user interactions. Users attempting to generate text reported failures and delays, with some noting that ChatGPT was not working. The outage began at Oct 22, 2025 8:41 PM UTC and was resolved by Oct 22, 2025 10:11 PM UTC, lasting about 1 hour and 30 minutes. The impact appeared global across the affected APIs rather than regional. OpenAI issued two status updates during the incident describing the elevated errors and ongoing mitigation efforts.

OpenAI's ChatGPT / Conversations experienced elevated error rates starting at Oct 22, 2025 2:22 PM UTC, with users reporting login failures and chats not loading or generating responses. The incident lasted about 19 minutes and was resolved by Oct 22, 2025 2:41 PM UTC. The issue affected the ChatGPT / Conversations component, causing degraded login and response performance for a subset of users.

OpenAI acknowledged the issue on their status page, noting elevated errors and ongoing mitigation, and posted 2 updates during the incident. The outage was brief and primarily impacted authentication and response generation for affected users, with service returning to normal after the updates.

OpenAI ChatGPT / Conversations experienced a brief outage with elevated error rates that prevented some users from receiving responses or loading chats. Reports from multiple regions described chats not loading, missing responses, or saved chats behaving abnormally. The incident began at 2:13 PM UTC and was resolved by 2:41 PM UTC, lasting about 28 minutes. OpenAI issued two status updates during the incident announcing elevated error rates and ongoing mitigation. StatusGator detected the issue at Oct 22, 2025 2:13 PM UTC, roughly 9 minutes before OpenAI's first acknowledgment.

OpenAI's API responses faced a brief degradation due to an elevated error rate, impacting the APIs / Responses component. The incident began at 23:51 UTC on 2025-10-21 and was resolved by 00:16 UTC on 2025-10-22, lasting about 25 minutes. The provider posted two status updates during the incident, noting that they were investigating the elevated error rate and subsequently confirming that services returned to normal.

The issues appeared to be limited to the API responses and did not indicate broader service outages. Users may have experienced errors or slower responses during the window, but service subsequently stabilized.

OpenAI experienced a Single Sign-On login disruption affecting sign-in to API services and ChatGPT. The issue prevented users from signing in via SSO for impacted components, specifically APIs / Login and ChatGPT / Login. The incident began at 8:43 AM UTC and was resolved by 10:11 AM UTC, lasting about 1 hour 28 minutes. OpenAI issued 2 status updates during the outage to communicate mitigation steps and restoration of access.

The incident was contained to the login flow and did not indicate a broader service impact beyond authentication, with full access restored after the updates.