OpenAI outage history

Users attempting to use data analysis and file creation features in ChatGPT encountered errors for a short period. The incident began at Oct 2, 2026 4:06 PM UTC and primarily affected the data analysis and file creation workflows, with some users reporting failures when running code or creating files. Affected component details on OpenAI’s status page pointed to the FedRAMP component as involved. One user even reported trouble loading conversations during the disruption, illustrating the kind of user experience impact this incident produced.

Impact began around 6:46 PM UTC with issues affecting login and signup for ChatGPT, followed by problems in Ads Manager and Ads API. Users reported sign-in failures, session drops, and errors when loading or using ChatGPT features and ad workflows, disrupting core authentication and content generation experiences across OpenAI services.

OpenAI issued status updates as the incident developed, with mitigation efforts noted around 7:06 PM UTC and 7:26 PM UTC. By 7:41 PM UTC, elevated errors were affecting ChatGPT Space Pages even as login and ads recovery progressed. The Space component returned to normal by 11:46 PM UTC, signaling widespread stabilization. Overall, the incident reflected broad impact on authentication and content-generation workflows, with users experiencing slower responses or timeouts during page creation or Codex interactions.

OpenAI experienced a broad disruption beginning around 5:00 PM UTC, with elevated latency across API endpoints and occasional timeouts. The impact soon extended beyond raw latency to authentication failures, with users reporting difficulties logging in and signing up. Across ChatGPT features—including ChatGPT Space pages and Codex workflows—users encountered slow responses, errors, and occasional 500 server errors, affecting users in multiple regions and driving a large volume of reports.

OpenAI experienced a broad outage that began in the early UTC hours, with elevated error rates for ChatGPT Pro and Plus and initial disruptions to ChatGPT / Conversations. The incident quickly expanded to ChatGPT Space Pages, causing widespread errors, timeouts, and degraded experiences for users worldwide who rely on core chat, page creation, and related features. By 10:36 AM UTC OpenAI indicated that all affected services had recovered and moved to monitoring mode, with ChatGPT / Conversations and ChatGPT / Space back up while vigilance remained in place.

An OpenAI outage began at 2:21 AM UTC with elevated errors in ChatGPT Space Pages, hindering page creation and live sessions. The issue expanded over the morning to elevated error rates for ChatGPT Pro and Plus users and degraded experiences in ChatGPT Conversations, with login problems affecting multiple features. OpenAI issued mitigations and began monitoring across affected components, including ChatGPT / Space, ChatGPT / Conversations, and related APIs such as Responses and Chat Completions.

User reports surged through the evening and into the next day, describing login failures, slow or unresponsive chats, and pages failing to load. The impact appeared broad across regions, with issues affecting workflows from chat to page creation and ads, and users encountered 500 errors and timeouts across services. OpenAI continued to publish updates detailing mitigations and observed improvements as services began to recover, though some components experienced slower performance during the recovery window.

OpenAI users faced a Help Center and live support chat outage beginning 5:51 PM UTC, with email support still available but no quick self-service guidance. The disruption hindered access to help and updates, while the broader platform remained partially usable for some, until the incident escalated.

As the incident evolved, elevated errors spread to ChatGPT, Codex, and the OpenAI APIs. Some users experienced failed requests, difficulty logging in or signing up, and tasks that would not complete. At around 6:27 PM UTC, OpenAI flagged elevated errors across multiple components, including ChatGPT features such as Conversations and GPTs, and API groups like Chat Completions, Responses, Embeddings, Images, Audio, Moderations, Realtime, and more, with the impact felt across login, file uploads, and related services.

Mitigation steps were applied by 7:41 PM UTC and monitoring continued through the evening. Support remained reachable by email, and recovery progressed over the next several hours. By 11:21 PM UTC, most components were reported up again, signaling a return to normal service after several hours of degraded performance across ChatGPT features and the OpenAI APIs. The incident appeared broadly felt in user reports around 18:00Z, reflecting widespread disruption across regions and use cases.

OpenAI Codex services experienced an outage that blocked access to Codex APIs, the CLI, Codex Web, and the VS Code extension. Users worldwide faced authentication errors and disrupted code-generation workflows as the issue unfolded. The incident began at 10:59 PM UTC and was resolved by 11:57 PM UTC.

OpenAI's website and ChatGPT experiences were not loading for users across multiple regions, with timeouts and the inability to reach the site reported starting around 12:45 AM UTC. The disruption appeared global, with reports from Europe, Asia, and the Americas, leaving many users unable to use ChatGPT or access the OpenAI site during the window.

StatusGator issued an Early Warning Signal, with issues detected at 1:53 AM UTC, roughly an hour before any official acknowledgment on OpenAI's status page. There were no published official status updates from OpenAI during the incident. Access was gradually restored and services returned to normal by 2:22 AM UTC.

Users began reporting elevated errors on GPT-6 Astra Pro around 6:15 PM UTC, with some unable to access features (a report from the Philippines described “Can’t Access anything”). OpenAI acknowledged elevated error rates at Sep 24, 2026 7:02 PM UTC and began implementing a mitigation.

Mitigation was applied by 7:36 PM UTC and monitoring continued as services recovered; access was restored by 7:46 PM UTC. The incident appeared to affect GPT-6 Astra Pro users for roughly an hour and a half before full recovery, with reports indicating degraded performance during that period.

Mobile users using OpenAI's ChatGPT experienced a UI issue where Work Mode and the Model Picker were not visible. This degraded the mobile chat experience, making it difficult to switch modes or select models during conversations. The problem appeared to affect the mobile interface broadly rather than a single feature, impacting a common workflow for users on mobile devices.

The issue began at 11:26 PM UTC and a mitigation was applied by 11:36 PM UTC, with full recovery by 11:52 PM UTC. Affected ChatGPT components were brought back online as the UI visibility was restored, and normal operation resumed for mobile users.