OpenAI Outage History

OpenAI experienced a brief partial outage affecting Codex Web task creation and the GitHub integration used by Deep Research. Users reported elevated errors when attempting to create new tasks in Codex Web, and some also encountered issues when connecting GitHub to Deep Research. The incident affected two components: ChatGPT / Deep Research and ChatGPT / Codex. The outage lasted from 9:06 PM UTC to 10:21 PM UTC, with a small volume of user reports during the event. OpenAI issued two status updates while addressing the issue.

OpenAI experienced a global outage impacting login and access to ChatGPT and related services. The issue began with degraded access at 12:16 PM UTC and escalated into a broader outage at 2:16 PM UTC, affecting three components: APIs / Login, Sora / Login, and ChatGPT / Login, with OpenAI attributing the disruption to an issue with a third-party service provider. User reports spanned multiple regions, indicating widespread impact on authentication and web access.

Recovery efforts proceeded with an initial fix by 2:51 PM UTC, followed by gradual improvement and full restoration across the affected components by 3:18 PM UTC. OpenAI issued several status updates during the incident to monitor the recovery. The high volume of user reports during the event reflected broad user impact across regions.

OpenAI ChatGPT experienced degraded performance for Plus users, with elevated error rates across ChatGPT Conversations and ChatGPT Agent interfaces. Users reported loading failures and error messages across ChatGPT features (e.g., infinite loading and attempts to load content). The incident began at 7:11 PM UTC when ChatGPT Conversations entered a degraded state, followed by elevated errors on ChatGPT Agent at 9:51 PM UTC. OpenAI identified the root cause and began mitigation by 12:41 AM UTC, with recovery progressing until the final component returned to normal by 7:51 AM UTC. The outage lasted roughly 12 hours and 40 minutes.

From Nov 14, 2025 8:56 PM UTC, OpenAI reported a subset of Batch API jobs were stuck in finalizing state, resulting in degraded performance for Batch operations and impacting ChatGPT file uploads. Users reported timeouts and delays in processing uploads and batch jobs. The incident persisted for about 12 hours, from Nov 14, 2025 8:56 PM UTC to Nov 15, 2025 8:56 AM UTC.

During the incident, OpenAI issued 3 status updates to communicate the investigation and remediation steps. By Nov 15, 2025 8:56 AM UTC, affected components including ChatGPT / File uploads and APIs / Batch were restored to normal operation.

OpenAI reported elevated error rates in the Fine-tuning API for gpt-4o-mini, causing degraded performance for fine-tuning tasks. The incident began at Nov 14, 2025 9:06 AM UTC and was resolved by Nov 14, 2025 9:41 AM UTC, lasting about 35 minutes. The issue affected the APIs / Fine-tuning component. The provider posted two status updates during the incident, first acknowledging the degraded state and then confirming recovery.

OpenAI experienced a multi-service outage starting at Nov 12, 2025 11:31 PM UTC that affected ChatGPT Conversations (Access Denied for enterprise users), ChatGPT Image Generation on Android, and the Login service. Users reported being unable to access conversations, image generation failures or timeouts, and login errors. The incident lasted about 1 day and 2 hours, with OpenAI posting 6 official updates as mitigation steps were applied and monitoring continued across the affected components. The outage began at Nov 12, 2025 11:31 PM UTC and was resolved by Nov 14, 2025 2:16 AM UTC.

OpenAI documented the progression from degraded service to recovery via six status updates, with the final update signaling recovery as the affected components returned to normal operation: ChatGPT / Conversations, ChatGPT / Image Generation, and ChatGPT / Login.

OpenAI experienced degraded performance in the Vector stores API, impacting vector-store operations. The issue began at 5:56 AM UTC and was resolved by 8:26 AM UTC, lasting about 2 hours and 30 minutes. OpenAI noted latency in processing the Vector stores API and implemented mitigations while monitoring recovery. The APIs / Files component was affected, and two status updates were issued during the incident. User reports were minimal during the event.

OpenAI experienced a partial outage starting at Nov 9, 2025 8:16 AM UTC when the Costs API endpoint (/v1/organization/cost) began returning 404 errors. The issue affected two API groups: APIs / Login and APIs / Chat Completions, which were degraded during the incident. User reports began around 7:15 AM UTC, indicating early visibility ahead of the acknowledgment. The outage lasted approximately 1 hour and 25 minutes, with the affected components returning to normal by Nov 9, 2025 9:41 AM UTC. The provider issued 2 status updates during the incident.

OpenAI experienced elevated error rates across the Sora API, affecting video generation and video viewing components, beginning at Nov 7, 2025 9:21 PM UTC. Users reported failures generating videos and playback issues, with errors appearing in multiple regions. The disruption also impacted Codex-related services (ChatGPT / Codex), signaling broader impact across OpenAI AI capabilities.

OpenAI's ChatGPT free tier experienced degraded availability starting around 6:12 PM UTC, with users seeing elevated errors affecting the ChatGPT / Conversations component. The OpenAI status page indicated a mitigation was being implemented, and the incident was listed as degraded starting at 6:26 PM UTC. The issue was resolved by 7:01 PM UTC, lasting about 49 minutes, and the provider posted 2 updates during the incident.