OpenAI’s ChatGPT and Sora were briefly down for users on Thursday, December 26, before coming back online. The outage affected users across the globe and lasted for hours, beginning at approximately 11 AM PT.
“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,” the AI firm wrote in a post on X. According to Downdetector – the outage tracking website – 88% of reports were about issues with ChatGPT, while 8% of reports revealed issues with OpenAI’s website.
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 (@OpenAI) December 26, 2024
The outage on Thursday was marked by widespread access issues for users trying to interact with ChatGPT, OpenAI’s AI-driven chatbot. Many reported being unable to load the chatbot or were greeted with error messages such as “internal server errors.” In other cases, users were able to load ChatGPT but found that their prompts went unprocessed. This was particularly disruptive for those who use the platform for both work and study purposes, as well as for developers who are reliant on OpenAI’s API to integrate it into their own applications.
The problem also affected Sora, OpenAI’s video generation platform, which had recently been made available to ChatGPT Plus subscribers. Initially, the platform experienced partial outages, but by 6:15 PM ET, OpenAI confirmed that Sora had returned to full functionality, though the recovery of ChatGPT and its API continued into the evening. “ChatGPT is mostly recovered and we are continuing to work on an overall fix,” the company wrote on its status page, adding that its APIs – which were affected by the service disruption as well – are operational once more.
OpenAI’s status page indicated that the root cause of the issue was linked to “high error rates” affecting its services, which stemmed from an issue with one of its upstream providers. However, the company did not provide specifics regarding which provider was responsible or the exact nature of the technical fault. Unsurprisingly, this led to widespread speculation online about the possible reasons for the outage, as users took to X (the microblogging platform) to air their grievances and expressing frustration over the downtime. The speculation includes theories related to OpenAI’s reliance on cloud infrastructure, specifically from tech behemoth Microsoft.
ChatGPT’s latest outage is the second time the chatbot experienced such a service disruption – something similar occurred on December 11 and lasted for approximately six hours. At that time, the company had attributed the disruption to a problem with a newly implemented telemetry service that had overtaxed the firm’s hardware resources.
