OpenAI Outage History

OpenAI's ChatGPT connectors and Codex experienced a multi-component disruption beginning on Aug 14, 2025 5:41 PM UTC. Plus and Pro users reported elevated errors when using the Use connectors tool, with the feature sometimes inaccessible for some users. The issue affected two components: ChatGPT / Context connectors and ChatGPT / Codex, and OpenAI issued four status updates as mitigation progressed.

From Aug 14, 2025 5:41 PM UTC to Aug 15, 2025 7:41 AM UTC, the outage lasted roughly 14 hours. The provider issued four status updates throughout the incident, detailing ongoing mitigation efforts and gradual restoration. Reports from users in diverse locations, including New Taipei City and Wilmington, Massachusetts, indicated that the issue affected users across regions, consistent with a broader, cross-regional disruption. By Aug 15, 2025 7:41 AM UTC, both affected components had returned to normal operation.

OpenAI's Compliance API experienced a brief degraded performance period affecting the APIs / Compliance component. The incident began at 6:41 PM UTC and was resolved by 7:31 PM UTC, lasting roughly 50 minutes. During this time, some users encountered errors when calling compliance-related endpoints. The OpenAI status page issued two updates describing the degraded service and subsequent recovery. User reports were limited, indicating only a handful of affected users.

OpenAI reported a brief degradation in the ChatGPT / Image Generation service, with image generation attempts experiencing errors for a short period. The issue affected the image generation feature used within ChatGPT, and OpenAI indicated they were investigating the incident as it unfolded.

OpenAI reported a brief image generation outage affecting free-tier ChatGPT users. Users attempting to generate images during the incident encountered errors, while other ChatGPT functions remained available. The issue began at 12:06 PM UTC and was resolved by 12:21 PM UTC, lasting about 15 minutes. The OpenAI status page described the incident as elevated errors in the Image Generation service and indicated that mitigation was being implemented.

OpenAI issued two status updates during the incident to communicate the impact and the mitigation progress, with the affected component listed as ChatGPT / Image Generation. The service returned to normal after the updates, with status returning to active for all users.

OpenAI experienced a partial outage affecting ChatGPT search features and several API endpoints. Users reported elevated errors and slower responses when triggering web-search actions in ChatGPT prompts, and Codex tasks also showed increased failure rates. The disruption touched multiple parts of the ChatGPT and API ecosystem, including the ChatGPT Search component, and API surfaces such as Responses, Assistants, Completions (legacy), and the Codex flow, with users experiencing degraded performance across these areas.

OpenAI's status page posted four updates at 3:56 PM UTC, 7:16 PM UTC, 8:01 PM UTC, and 10:07 PM UTC as engineers worked to mitigate. The outage began at 3:56 PM UTC and was resolved by 10:07 PM UTC, lasting roughly 6 hours 11 minutes. A small volume of user reports accompanied the incident, reflecting the partial nature of the disruption.

OpenAI ChatGPT Conversations experienced a brief degradation due to GPT-5 rate limit and model-not-found errors, impacting user sessions in the ChatGPT / Conversations component. The incident lasted about 5 minutes, from Aug 8, 2025 7:11 PM UTC to Aug 8, 2025 7:16 PM UTC. OpenAI mitigated the issue and monitored the recovery, issuing two status updates as the service returned to normal.

OpenAI reported a partial outage of the ChatGPT Search feature affecting free ChatGPT users beginning at Aug 8, 2025 10:46 AM UTC UTC, with search queries not returning results or performing slowly. The issue appeared to be global, impacting users in multiple regions who rely on ChatGPT Search to locate information. The outage persisted for several hours as OpenAI investigated and issued updates, with two status updates posted during the incident. The service began returning to normal and the status was restored at Aug 8, 2025 6:11 PM UTC UTC.

The outage began on Aug 7, 2025 12:31 PM UTC with elevated errors on the ChatGPT 4.5 model. Later, at Aug 7, 2025 8:16 PM UTC, users experienced degraded latency and higher error rates for file uploads to vector stores, affecting three OpenAI components: ChatGPT / Conversations, APIs / Assistants, and APIs / Files. Users could encounter failed or slower ChatGPT interactions and delays uploading files for vector storage. The incident persisted for roughly 21 hours and was resolved by Aug 8, 2025 9:06 AM UTC.

OpenAI issued four status updates during the incident and implemented mitigations; vector store components were restored by Aug 8, 2025 12:46 AM UTC and ChatGPT / Conversations recovered by Aug 8, 2025 9:06 AM UTC.

The OpenAI ChatGPT Deep Research feature experienced elevated error rates for free-tier users, causing intermittent failures and degraded access to Deep Research. The issue was isolated to the Deep Research component and did not indicate a full outage across OpenAI services. User reports were limited, with some users encountering errors while others could continue normally.

OpenAI's status page shows the issue moved to a warn state at 10:41 AM UTC and returned to normal at 12:06 PM UTC, after two official updates were issued during the incident. The outage lasted approximately 1 hour 25 minutes.

OpenAI's ChatGPT Conversations experienced elevated errors affecting image-based chats, leading to failed responses and contextless errors reported by users across multiple regions. The outage began on Aug 6, 2025 8:41 PM UTC and was resolved by Aug 6, 2025 10:01 PM UTC, lasting about 1 hour 20 minutes. OpenAI documented elevated error rates and ongoing mitigation during the incident, followed by restoration to normal operation.

OpenAI issued two status updates during the incident. The first acknowledged elevated error rates with mitigation in progress, and the second confirmed that the service had returned to normal. The issue impacted the ChatGPT Conversations component and produced a degraded experience for users attempting image-based conversations worldwide.