OpenAI Outage History

OpenAI's ChatGPT experienced a brief global disruption with elevated errors across many features, making chats, prompts, and related capabilities intermittently unusable for about 20 minutes. The incident impacted 13 core components including ChatGPT / Codex, Voice mode, GPTs, Agent, ChatGPT Atlas, Connectors/Apps, Search, Conversations, Image Generation, Login, Compliance API, File uploads, and Deep Research. Users worldwide reported issues such as chats not loading, prompts failing, and various errors during the incident.

StatusGator detected the issue early at Feb 4, 2026 5:15 PM UTC, roughly 26 minutes before OpenAI acknowledged the problem. OpenAI issued 2 status updates during the incident and services were restored by 6:01 PM UTC. The outage lasted from 5:41 PM UTC to 6:01 PM UTC. A high volume of user reports indicated broad, worldwide impact across users.

OpenAI experienced a brief degradation of ChatGPT / Conversations, with some users reporting that they could not load or start conversations. The issue began at 1:26 AM UTC and lasted about 2 hours and 20 minutes, during which the ChatGPT / Conversations component was listed as degraded. OpenAI said they were investigating the problem and would provide updates as they worked toward a fix.

OpenAI issued two status updates during the incident, and the service returned to normal at 3:46 AM UTC. A sample user report from Los Angeles around 02:42 UTC described trouble generating responses, illustrating the symptom reported by affected users. Overall, the impact appeared limited to a subset of users rather than a full service outage.

OpenAI's ChatGPT services experienced elevated error rates that impacted login, conversations, and a broad set of APIs, leading to widespread disruption for users trying to chat, generate responses, or access their chats. The outage began around 7:45 PM UTC with initial user reports, and StatusGator detected the issue at Feb 3, 2026 7:54 PM UTC, about 42 minutes before the first official update at Feb 3, 2026 8:36 PM UTC. Affected components included ChatGPT Atlas, Conversations, Login, Deep Research, Agent, Codex, Compliance API, Search, File uploads, Voice mode, GPTs, Image Generation, Connectors/Apps, APIs / Login, and Fine-tuning. The incident involved a large volume of user reports at the peak, reflecting global disruption.

OpenAI reported a brief outage affecting sign-up for ChatGPT when using Google Authentication on Android. The issue prevented some users from completing onboarding with Google Sign-In, impacting the ChatGPT login flow. The outage began at 8:41 PM UTC and was resolved by 9:16 PM UTC, lasting about 35 minutes. The OpenAI status page provided mitigation updates as the issue was addressed, and normal sign-up resumed. The incident involved the ChatGPT / Login component.

The OpenAI ChatGPT / Codex service experienced degraded performance and occasional timeouts for users worldwide. The incident began on Jan 27, 2026 9:06 AM UTC when OpenAI listed the ChatGPT / Codex component as degraded due to an elevated Codex error rate. User reports from Manchester, Lisbon, Utrecht, and Hong Kong described slow responses and timeouts. The outage lasted approximately 2 hours and was resolved by Jan 27, 2026 11:06 AM UTC.

OpenAI issued two status updates during the incident, documenting the degraded condition and later the resolution as services returned to normal. The issues affected ChatGPT and Codex components globally, indicating a broad service disruption rather than a regional outage.

OpenAI experienced a brief issue where Connectors/Apps could not be selected in the ChatGPT Input Menu, temporarily disrupting workflows that rely on Connectors/Apps. The incident began at Jan 12, 2026 7:37 PM UTC when the status moved to warning, and was resolved by Jan 12, 2026 7:46 PM UTC after about 9 minutes. OpenAI issued 2 status updates during the incident, describing the ongoing investigation and the eventual restoration of normal operation. The issue was localized to the ChatGPT Connectors component and did not indicate broader service impact.

OpenAI experienced degraded performance for image-containing prompts in ChatGPT conversations and related API endpoints. Users submitting prompts with images encountered errors or failed responses across ChatGPT Conversations, APIs / Chat Completions, and APIs / Responses. The issue began at Jan 8, 2026 8:26 PM UTC and persisted for roughly 55 minutes before services returned to normal at Jan 8, 2026 9:21 PM UTC. OpenAI issued 2 status updates during the incident, describing the elevated error rates and subsequent restoration of service.

The impact was focused on image-based prompts across three components and was resolved once the underlying issue was addressed.

OpenAI reported a brief degradation affecting ChatGPT / Codex cloud tasks, with some users experiencing failures to load replies or slower processing. The issue began at Jan 8, 2026 2:16 AM UTC and was resolved by Jan 8, 2026 2:31 AM UTC, lasting about 15 minutes. The OpenAI status page issued two updates during the incident, noting that some Codex cloud tasks were failing and that the team was investigating, before restoring normal service. A user in the United Kingdom reported not loading replies during the event.

OpenAI experienced elevated error rates on the Responses API, resulting in degraded performance for requests to the API. The incident affected the OpenAI Responses component and lasted from Jan 7, 2026 11:06 PM UTC to Jan 7, 2026 11:56 PM UTC. Users may have seen failed requests or slower responses during this window.

OpenAI's status page issued two updates during the incident. The initial update indicated elevated errors and that mitigation was being implemented; the follow-up confirmed resolution and return to normal operation. The outage duration was approximately 50 minutes, with the Responses API returning to normal by Jan 7, 2026 11:56 PM UTC.

OpenAI's ChatGPT experienced sign-in authentication problems and intermittent loading for users worldwide after SSO. Users reported sign-in errors and blank sign-in screens beginning around 9:55 PM UTC and continuing through the evening, with the early warning signal ending at 10:27 PM UTC. StatusGator detected the issue at Dec 18, 2025 9:55 PM UTC, about 1h26m before OpenAI officially acknowledged it at Dec 18, 2025 11:21 PM UTC.