OpenAI Outage History

The incident began at Nov 26, 2025 12:06 AM UTC when OpenAI flagged an increase in ChatGPT errors for some users. A broad set of ChatGPT components experienced degraded performance, impacting various features such as ChatGPT Conversations, Search, File uploads, Voice mode, Deep Research, Agent, Codex, and related tools. Users reported intermittent issues and reduced reliability across multiple capabilities during the event.

OpenAI issued two status updates during the incident; by Nov 26, 2025 3:06 AM UTC the affected components had returned to normal, marking an incident duration of about 3 hours. The updates confirm ongoing investigation followed by recovery across multiple ChatGPT features.

OpenAI began investigating ImageGen issues affecting ChatGPT Workspaces with EU Data Residency at 7:41 PM UTC, followed by a broader alert for Workspaces with Non-US Data Residency at 8:16 PM UTC. The Image Generation component experienced degraded performance and, in some cases, disappearing generated images for users. The incident prompted multiple status updates as the team diagnosed the root cause and implemented a fix. The outage lasted about 3 hours, beginning 7:41 PM UTC and ending with the Image Generation service restored at 10:41 PM UTC. OpenAI issued 4 updates during the incident at the times noted above. User reports described image loss or disappearing results, including a report from Chicago at 10:38 PM UTC.

OpenAI reported elevated errors impacting ChatGPT / Codex, causing degraded performance for Codex-related features. The incident began on Nov 23, 2025 4:26 AM UTC and was resolved by Nov 23, 2025 6:56 AM UTC, lasting about 2 hours and 30 minutes. OpenAI issued an initial status update noting the elevated errors and a follow-up after applying mitigation while monitoring recovery.

OpenAI's ChatGPT and Codex experienced elevated errors beginning at Nov 22, 2025 7:21 PM UTC, leading to degraded responsiveness and errors for users across Codex and related ChatGPT features. Reports described non-responsive Codex endpoints and other interaction failures, with activity observed from multiple U.S. regions (e.g., Tennessee, Utah, Hawaii) over several hours. The incident did not indicate a full outage but caused noticeable disruption for developers and end users relying on Codex and ChatGPT tooling.

The provider issued two status updates during the incident, describing elevated errors and ongoing investigation, and services were restored by Nov 22, 2025 10:51 PM UTC. The outage lasted roughly 3 hours and 30 minutes, with a return to normal operation after the final update. Reports from several U.S. locations suggest a regional impact rather than a global one.

OpenAI experienced a brief degradation affecting Codex Web task creation and certain Deep Research connections (including GitHub integration). Users reported errors when creating new tasks in Codex Web and intermittent issues with GitHub links through Deep Research. The incident impacted two OpenAI components: ChatGPT / Codex and ChatGPT / Deep Research, with affected users experiencing task creation failures and related API problems. The outage began at Nov 18, 2025 9:06 PM UTC and was resolved by Nov 18, 2025 10:21 PM UTC, lasting about 1 hour and 15 minutes. The OpenAI status page documented the issue as a degradation affecting these features, followed by a restoration once mitigation was in place. During the incident, OpenAI issued two status updates outlining the degraded task creation and GitHub integration issues and noting that a mitigation was being implemented; later updates indicated the service had returned to normal. User reports described API errors and related disruptions consistent with the described symptoms.

OpenAI users experienced login failures across the platform starting at 2:18 PM UTC. Many reported being unable to sign in or access account features, with some facing Cloudflare challenge pages, and others seeing 504 errors. The disruption lasted until 2:47 PM UTC. StatusGator logged an Early Warning Signal at Nov 18, 2025 2:18 PM UTC, indicating the issue was detected before OpenAI's status updates were published. The surge in user reports around mid-afternoon indicates a broad impact on authentication and login workflows across OpenAI services.

The outage began on 2:16 PM UTC and was resolved by 3:18 PM UTC. OpenAI experienced a global login and access disruption affecting ChatGPT and OpenAI websites, including APIs / Login, ChatGPT / Login, and Sora / Login. Users reportedSign-in failures, page load errors, and authentication problems across regions, with some users seeing common error messages and downtime indicators. The incident progressed through multiple status changes, starting with a warn designation, moving to a full outage, and then returning to degraded states before recovery. A third-party service provider issue was identified as the root cause, and OpenAI issued updates as the situation evolved.

OpenAI ChatGPT experienced a degraded performance event beginning at Nov 17, 2025 7:11 PM UTC with elevated error rates affecting ChatGPT for Plus users and degraded performance for the ChatGPT Agent and Conversations components. Users reported a range of symptoms, including errors, content loading failures, and persistent messages, across regions, indicating a broad impact. The incident lasted until Nov 18, 2025 7:51 AM UTC, a duration of about 12 hours and 40 minutes. The Agent component had recovered by 12:41 AM UTC, with Conversations returning to normal by 7:51 AM UTC.

OpenAI experienced a degraded state as a subset of Batch API jobs became stuck in the finalizing state, delaying batch processing and impacting related ChatGPT file uploads. The issue began at Nov 14, 2025 8:56 PM UTC and lasted about 12 hours, with the services returning to normal by Nov 15, 2025 8:56 AM UTC. Users reported failures and timeouts in batch workflows, and the impact was seen across two components: APIs / Batch and ChatGPT / File uploads.

OpenAI experienced a brief degradation in the Fine-Tuning API for the gpt-4o-mini model, with elevated error rates potentially affecting fine-tuning tasks. The issue began at Nov 14, 2025 9:06 AM UTC and was resolved by Nov 14, 2025 9:41 AM UTC, lasting about 35 minutes. OpenAI issued two status updates documenting an investigation and the subsequent restoration of the service. The impact was limited to the Fine-Tuning API, with other OpenAI services unaffected.