ChatGPT goes down two days running: Friday's 30-minute outage, Saturday morning latency hit
ChatGPT went down Friday evening — around 30 minutes offline, 19,403 Downdetector reports — then hit elevated latency again Saturday morning, resolved by 10:31 a.m. PT. Altman owned it on X: much better reliability than before, clearly more work ahead.
What happened
Friday evening, ChatGPT went down hard enough to draw more than 19,403 Downdetector reports, peaking at 7:13 p.m. ET. OpenAI said most users were back by 4:34 p.m. PT, with a few still reporting problems until 5 p.m. PT. Sam Altman said on X the chatbot had been offline for around 30 minutes: "We are much, much better at reliability than we used to be, but clearly more work in front of us," accompanied by a frown emoji. OpenAI's own announcement read simply: "Issues with ChatGPT have now been resolved."
Saturday morning brought a second hit. OpenAI's status page showed it investigating elevated latency disrupting ChatGPT and the API; by 8:15 a.m. PT the incident was active. Users saw blank chat windows or endlessly loading screens, with some users in India having particular trouble with chats and coding features. At 9:15 a.m. PT the ChatGPT issue was reported resolved while the API slowdown lingered, and by 10:31 a.m. PT the issue appeared resolved.
Why it matters
Two disruptions in about 18 hours, weeks after the DevDay 2026 traffic surge and the Dots rollout — scale keeps testing OpenAI's reliability promises in public. Altman's own framing concedes the gap: better than before, still not there.
Verified October 17, 2026
Sources
Get updates like this every morning
- ① Email
- ② Card on Stripe
- ③ 7 days free
Then $2/month · cancel anytime in one click