At around 11:00 a.m. PT / 2:00 p.m. ET on Wednesday, December 26th, ChatGPT went down for users worldwide. As OpenAI noted on its status page, users started experiencing high error rates on ChatGPT, its API, and Sora. After a short investigation period, OpenAI determined the issue was being caused by an upstream provider.
The status page shows that ChatGPT and the company’s API services are experiencing what it calls a “major outage,” while Sora is dealing with a “partial outage.”
This is not the first outage that OpenAI has dealt with in recent weeks, but it is among the most serious. On Friday, December 20th, users experienced elevated error rates with Dall-E and the Moderation API. One day earlier, users were also failing to sign up using their usernames and passwords or receive verification emails when logging into ChatGPT.
According to Downdetector, reports began spiking around 10:15 a.m. PT. The good news is that those reports took a precipitous drop after 11:00 a.m. PT, which is when the official investigation kicked off. As of 12:51 p.m. PT, I’m still unable to use ChatGPT.
“Something went wrong,” the chatbot explained. “If this issue persists please contact us through our help center at help.openai.com.”
We’ll update this post when OpenAI shares more information.
