OpenAI’s ChatGPT is back up and running after glitches lead to “elevated error rates” for those using the generative artificial intelligence chatbot.
The first OpenAI post came just before 4am on the West Coast of the USA: “We are currently experiencing elevated error rates in the API [Application Programming Interface]. We are currently investigating.”
Half an hour later, joy: “The issue has been identified and a fix is being implemented … we are monitoring the results.”
Thirteen minutes on, an update: “This issue has now been resolved. Between 3:33 AM and 4:23 AM PST, customers experienced elevated error rates on ChatGPT and the API.”
At 5:12AM, 29 minutes after the all-clear was given, more bad news.

OpenAI ChatGPT back on the air.
“We are currently experiencing elevated error rates for ChatGPT,” OpenAI said in a post. “We are currently investigating.”
On social media, people who have become attached to the large language model bot were expressing concern about how they would cope.
And then, at 5:43AM, a glimmer of hope: “We have identified the root cause of this issue, and are currently working to implement a fix.”
The minutes ticked by. At 6:34AM: “We are continuing to work on a fix for this issue.”
Thirty-five minutes later, at 7:09AM, champagne corks popped: “A fix has been implemented and we are monitoring the results.”
And at 8:05AM, just as millions of Californians were heading off to work, closure. For now. “This issue has now been resolved. Between 4:23am and 7:10am PST, customers experienced elevated errors on ChatGPT.”
The day before OpenAI posted that “some users may experience login issues on platform.openai.com”.