ChatGPT and Sora Outage: Services Restored After Hours of Downtime

OpenAI's popular AI platforms, ChatGPT and Sora, experienced a significant outage on Thursday, December 26, leaving users unable to access services for hours. The disruption began around 11 AM PT and was resolved later in the day, sparking frustration among users worldwide.

"Most of ChatGPT, the API, and Sora have been down for a couple of hours, and we're sorry for the trouble this is causing. We've identified the issue and have started recovery. We hope to be back asap," OpenAI posted on X (formerly Twitter) during the outage.

Downdetector revealed that 88% of the reports were related to ChatGPT access issues, while 8% pointed to problems with OpenAI's website. Users reported encountering error messages like "internal server errors" or experiencing unprocessed prompts. Some managed to load the chatbot, but its functionalities remained unresponsive. The outage affected not only individuals relying on ChatGPT for work and studies but also developers integrating its API into applications.

Sora, OpenAI's video generation platform for ChatGPT Plus subscribers, was also impacted. Initially experiencing partial disruptions, it was fully operational by 6:15 PM ET. However, the recovery process for ChatGPT and its API extended into the evening. OpenAI later confirmed that services had largely been restored, stating, "ChatGPT is mostly recovered, and we are continuing to work on an overall fix."

OpenAI's status page identified the root cause of the issue as "high error rates" linked to one of its upstream providers, though specific details were not disclosed. This lack of clarity prompted speculation online, with some attributing the outage to OpenAI's reliance on cloud infrastructure, particularly Microsoft's services. Users took to X to express frustrations and theorize about the technical fault.

This incident marked the second significant disruption for ChatGPT in December. Earlier in the month, a six-hour outage occurred due to issues with a telemetry service overloading OpenAI's hardware resources. The December 26 outage underscored the challenges of maintaining reliability for high-demand AI services. OpenAI continues working to address these challenges while ensuring minimal disruptions for its global user base.