OpenAI Outage History

OpenAI reported an increased error rate for ChatGPT, affecting two components: ChatGPT / Codex and ChatGPT / Agent. Users may have experienced errors or degraded responses during the incident. The outage began at 10:01 PM UTC and lasted about 3 hours, ending by 1:01 AM UTC. The OpenAI status page issued 2 updates during the incident.

OpenAI's ChatGPT file uploads experienced elevated errors, causing degraded performance for users attempting to upload files. The incident began at Sep 22, 2025 4:21 AM UTC and lasted 5 hours 35 minutes, ending at 9:56 AM UTC. The issue affected the ChatGPT / File uploads component, with users potentially seeing upload failures or delays during the period. OpenAI issued two official status updates as mitigation measures were implemented, and service returned to normal after the final update.

OpenAI experienced an outage affecting the ability to create new projects across ChatGPT Web and mobile apps. The issue impacted ChatGPT / Web, ChatGPT / Mobile - iOS, ChatGPT / Mobile - Android, and later also affected file uploads. The outage began at 7:56 PM UTC and lasted until 11:12 PM UTC, a total duration of about 3 hours and 16 minutes. During the incident, the provider issued three status updates detailing investigation, mitigation steps, and restoration actions.

An outage affected the ChatGPT file uploads feature, preventing users from attaching files to chats or projects and disrupting workflows that rely on file submissions. Users reported errors when attempting to upload and encountered issues described in reports across several regions.

The OpenAI status page reported elevated error rates affecting ChatGPT Agent and GPT-5 Pro, with ChatGPT Conversations also listed as degraded. Users interacting with these features may have experienced errors or slower responses. The outage began at 4:51 PM UTC and was resolved by 6:31 PM UTC, lasting roughly 1 hour and 40 minutes. Affected components included ChatGPT / Agent and ChatGPT / Conversations.

OpenAI issued 3 status updates during the incident as investigations progressed. User reports were limited, with at least one observation around 6:00 PM UTC describing issues with project files not uploading or deleting.

OpenAI's Conversations with Images for Free Users experienced elevated errors, causing interruptions for image-enabled chats. The disruption began at Sep 17, 2025 10:07 PM UTC and was resolved by Sep 17, 2025 10:41 PM UTC, lasting about 34 minutes. The impact appeared isolated to this feature with no broader OpenAI services indicated as affected. OpenAI noted work on a mitigation and posted two status updates during the incident.

OpenAI's updates show the component briefly operated in a degraded state before service returned to normal by the end of the event. The incident involved limited scope but affected users attempting image-enabled conversations on Free Accounts.

OpenAI experienced a period of degraded performance affecting the GPT Store and single sign-on (SSO) login workflows. The incident began at Sep 16, 2025 3:46 PM UTC and affected multiple components, including ChatGPT / GPTs, Sora / Login, APIs / Login, and ChatGPT / Login, leading to elevated errors, sign-in failures, and delays in generating responses for users worldwide. The outage lasted about 25 hours and 50 minutes and was resolved by Sep 17, 2025 5:36 PM UTC. A moderate volume of user reports described issues ranging from login problems to unresponsive chat interactions.

The outage began on Sep 11, 2025 1:11 AM UTC and was resolved by Sep 16, 2025 4:51 AM UTC. OpenAI reported degraded performance across multiple API surfaces, with retrieval quality issues in file_search and elevated errors in several services. Users worldwide experienced slow responses, timeouts, and intermittent failures when using Assistants, Responses, Fine-tuning, and related APIs. The incident impacted 12+ components across the API surface, including Embeddings, Batch, Files, Images, Audio, Realtime, Login, Fine-tuning, Completions (legacy), Chat, Compliance, Moderations, and Uploads.

OpenAI reported a brief degradation affecting the APIs / Batch component, with users experiencing elevated errors when calling the Batch API. The outage began at Sep 11, 2025 12:46 AM UTC and was resolved by 12:51 AM UTC, lasting roughly 5 minutes. OpenAI issued 2 status updates during the incident, noting the elevated errors and the mitigation that restored normal operation.

OpenAI's ChatGPT / Agent experienced elevated error rates beginning at 7:02 PM UTC, resulting in degraded performance for some users. Users attempting to use ChatGPT features may have encountered errors or slower responses. A report from Reston, Virginia around 7:13 PM UTC indicates the issue was affecting users in at least one location. The incident prompted mitigation efforts from OpenAI as they investigated the cause.