OpenAI Outage History

OpenAI experienced a memory-related incident that impacted ChatGPT memory and Conversations. Users reported missing memories and errors affecting memory-related interactions, with several sessions failing or returning error messages. The issue began at 6:46 PM UTC and was resolved by 4:16 PM UTC, lasting roughly 21 hours and 30 minutes. The ChatGPT / Conversations component was identified as affected during the incident.

OpenAI ChatGPT Conversations experienced degraded performance due to an uptick in long-conversation errors. The incident began at 4:11 AM UTC when OpenAI flagged that users were encountering 'conversation too long' messages more frequently than expected. The issue affected the ChatGPT / Conversations component and appeared to impact users attempting extended chats. The outage lasted about 1 hour and 10 minutes, with resolution at 5:21 AM UTC.

A small set of user reports described the symptoms during the incident, and OpenAI issued two status updates as the situation progressed, before returning to normal operation.

OpenAI's ChatGPT and Codex experienced elevated errors affecting user interactions. Users reported issues generating content and saw error prompts such as “try again later.” The incident began at Nov 5, 2025 7:46 PM UTC and was resolved by Nov 5, 2025 9:56 PM UTC, lasting roughly 2 hours and 10 minutes. The provider’s status page described an Elevated Error and noted that mitigation was being implemented, with the affected component listed as ChatGPT / Codex.

OpenAI reported a login issue across multiple services, blocking sign-in for current and new users of APIs / Login, Sora / Login, and ChatGPT / Login. Some users described login failures and 500 errors, preventing access to the platform or API keys.

The incident was acknowledged on OpenAI's status page and mitigation efforts were underway. The outage began at Nov 4, 2025 3:02 PM UTC and was resolved by Nov 4, 2025 5:41 PM UTC, lasting about 2 hours and 39 minutes. The provider issued two status updates during the incident, detailing ongoing mitigation and the eventual restoration of login services.

OpenAI reported degraded availability for Codex on the Web, impacting Codex Tasks in the browser. From 19:15 to 20:15 Pacific Time, users were unable to run Codex Tasks on the web. OpenAI applied a mitigation and is monitoring the fix as the service returns toward normal. This disturbance involved the ChatGPT / Codex component.

The incident spanned roughly one hour, during which two official status updates were published. The first update, indicating degraded availability, was posted at Nov 4, 2025 4:26 AM UTC and the second confirmed restoration at Nov 4, 2025 5:16 AM UTC. The service was fully restored by then, and OpenAI resumed normal operations.

OpenAI's Responses API experienced a brief degradation with elevated errors observed by users. The incident began at Nov 4, 2025 2:06 AM UTC when the component moved to a degraded state and was resolved by Nov 4, 2025 2:31 AM UTC, lasting about 25 minutes. The issue affected the APIs / Responses component, and users may have encountered error responses during that window. The provider posted two updates during the incident, and mitigations were implemented to restore normal service.

OpenAI's Responses API operating in background mode experienced degraded performance with elevated latency and timeouts, affecting applications that rely on background processing. User reports began around 12:45 PM UTC, including a sample from Porto Alegre describing inability to send photos and text, indicating a broader impact on interactions with the API.

The incident began with the status change to warn at 1:06 PM UTC and was resolved by 6:56 PM UTC, lasting about 5 hours and 50 minutes. OpenAI issued two status updates during the incident, clarifying the degraded performance and the eventual restoration. The affected component was APIs / Responses.

The incident began at Oct 30, 2025 7:11 PM UTC when OpenAI flagged ChatGPT / Conversations as degraded while they investigated the issue. Users worldwide reported errors and unresponsive conversations, indicating a broad impact on ChatGPT Conversations. The reports came from many regions suggesting global reach rather than a regional outage.

OpenAI acknowledged the issue and began investigating, with the status page noting 'Increased errors on ChatGPT conversations' and that they were investigating the issue for the listed services. The provider issued 2 status updates during the incident before the service returned to normal at Oct 30, 2025 9:21 PM UTC. The outage lasted approximately 2 hours 10 minutes, with a surge in user reports around the peak period across multiple locations.

Some OpenAI users were unable to activate ChatGPT Connectors that were already enabled in their workspace. Although connectors appeared available in Workspace Settings, activation could not be completed for their group or role. The outage lasted approximately 1 hour and 30 minutes, beginning at Oct 29, 2025 10:46 AM UTC and ending at Oct 29, 2025 12:16 PM UTC.

OpenAI documented the issue on the status page and issued two updates as engineers investigated and then restored functionality. The ChatGPT / Connectors component transitioned from a degraded state back to normal once the issue was resolved.

OpenAI reported a brief outage affecting ChatGPT conversations for paid users. The incident began at Oct 27, 2025 3:31 PM UTC when the status page moved from up to warn, citing increased errors in ChatGPT conversations and an active investigation. The issue affected the ChatGPT / Conversations component, causing users to see errors or failed message streams when generating outputs. The outage lasted about 41 minutes and was resolved by Oct 27, 2025 4:12 PM UTC. OpenAI issued two status updates during the incident as they investigated and then restored service. User reports indicate a global distribution of impact, with observations from cities across North America, Europe, Asia, and Australia reporting errors such as failed message streams and retry loops.