OpenAI Outage History

OpenAI reported a brief degraded state affecting ChatGPT login starting at Apr 1, 2025 11:01 AM UTC. The issue was listed as ChatGPT / Login in a degraded state and then returned to normal by Apr 1, 2025 12:31 PM UTC. The provider issued 2 status updates during the incident, and the outage duration was approximately 1 hour and 30 minutes.

During this period, users attempting to sign in to ChatGPT may have experienced authentication delays or login failures. The impact was isolated to the ChatGPT login component.

The outage began at 12:36 PM UTC and affected a broad set of OpenAI services, including ChatGPT login, ChatGPT Web, Playground, Labs, and related features, as well as API responses. Users worldwide reported slow responses, login failures, and errors (e.g., 500/503) across ChatGPT and API interactions, with impact spanning North America, Europe, and Asia.

OpenAI ChatGPT / Web experienced a brief degraded performance affecting user sessions due to token-related errors. The issue began at 10:30 AM UTC and impacted users across devices, with authentication failures reported. The status page listed a degraded state beginning at 11:11 AM UTC and the service was restored by 11:43 AM UTC, spanning roughly 1 hour and 13 minutes.

The provider posted two official status updates during the incident, acknowledging the degraded performance and later confirming restoration. A sample user report cited errors such as "could not find user for provided access token" during the outage.

OpenAI ChatGPT Web users faced login failures and token related errors, preventing authentication and access to ChatGPT Web across regions. Reports described errors such as could not find user for provided access token and token_exchange_failed, indicating problems with the authentication flow and session creation. The incident began with login issues at 10:41 AM UTC and affected users across multiple regions including Australia, the United Kingdom, Germany, Croatia, and Uruguay, with symptoms observed on various devices.

OpenAI acknowledged a degraded service on their status page, with updates at Mar 30, 2025 11:11 AM UTC (up to warn) and Mar 30, 2025 11:43 AM UTC (warn to up). StatusGator surfaced an Early Warning Signal beginning at Mar 30, 2025 10:41 AM UTC, ending at 12:01 PM UTC, indicating that users reported issues before the official acknowledgment. The main degradation lasted roughly from 11:11 to 11:43 UTC, and the service was back to normal by 11:43 UTC, with some residual reports continuing until the early warning signal ended. The incident involved the ChatGPT Web component and appeared as a global authentication disruption, with two official status updates during the incident.

OpenAI experienced degraded performance affecting multiple components, including ChatGPT File uploads, ChatGPT Image Generation, and the APIs for Images and Files. The issue began at 1:41 PM UTC and was resolved by 3:37 PM UTC, lasting roughly 1 hour and 56 minutes. User reports started around 2:00 PM UTC and continued through 2:45 PM UTC, with users seeing errors or slower responses when uploading files, generating images, or accessing image and file APIs. The impact appeared to be widespread across ChatGPT features and API endpoints, though not a full outage.

OpenAI's status page documented the incident with four updates, tracking the issue from initial warning to recovery as components were brought back online. By late afternoon, all previously degraded services were reported as operational again.

OpenAI Sora experienced a brief degraded performance affecting five components: Image Generation, Login, Video generation, Video viewing, and Feed. The degradation began at Mar 28, 2025 4:16 PM UTC and was resolved by Mar 28, 2025 8:02 PM UTC, lasting about 3 hours and 46 minutes. OpenAI issued two status updates during the incident to communicate the issue and restoration.

OpenAI reported a brief degradation impacting two API components: APIs / Chat and APIs / Responses. The incident began at 8:56 PM UTC and was resolved by 10:06 PM UTC, lasting about 1 hour and 10 minutes. Users of the Chat and Responses endpoints may have experienced slower responses or intermittent issues during the outage. The provider issued 3 status updates describing the degraded state and the subsequent recovery.

From Mar 26, 2025 9:42 PM UTC to Mar 27, 2025 1:46 AM UTC, OpenAI's ChatGPT and Search experienced degraded performance, with slower responses and intermittent results reported by users attempting to use those features. The outage lasted approximately 4 hours and 4 minutes and was confined to the ChatGPT / Search component, with no other OpenAI services indicated as affected.

The provider's status page documented the degradation and published two official updates during the incident before the service returned to normal at Mar 27, 2025 1:46 AM UTC.

OpenAI experienced widespread degraded performance affecting ChatGPT web and mobile clients and a broad API suite starting at 3:16 PM UTC. Users encountered errors and slower responses across ChatGPT Web, iOS, and Android, while API users faced intermittent failures across multiple endpoints. At the incident peak, about a dozen API components were degraded, including Chat, Embeddings, Batch, Fine-tuning, Assistants, Realtime, Images, Responses, Audio, Files, Completions (legacy), Compliance, Moderations, and Uploads. The issue persisted for roughly 10 hours and 30 minutes, with users reporting intermittent problems during this period. By 1:46 AM UTC the Chat API component had returned to normal and overall stabilization followed as status updates continued.

The outage began at 2:31 PM UTC and was resolved by 3:47 PM UTC. OpenAI ChatGPT web and mobile experiences experienced degraded performance during this interval, with users on the Web, iOS, and Android reporting slower responses and intermittent errors. The outage lasted about 1 hour 16 minutes.

OpenAI's status page communicated the incident with two updates, noting degraded performance for the affected components and then restoration of normal service after the remediation. The incident highlights the temporary disruption to ChatGPT across web and mobile platforms.