ChatGPT's paid plans saw elevated error rates for 1h 58m on 4 November 2026, starting at 22:38 UTC. The fault sat on OpenAI's side, affecting the Web and ChatGPT components.
Started
22:38 UTC
Duration
Lasted 1h 58m
Source
IsDown
Next time
OpenAI is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
Affected components: Web, ChatGPT
What happened?
OpenAI logged a single incident on this day, starting at 22:38 UTC under the title "Elevated error rates across paid ChatGPT plans". The first update said some users on paid ChatGPT plans might see errors when sending messages, and that a mitigation was being worked on. Eleven minutes later, at 22:49 UTC, OpenAI posted that a mitigation had been applied and that they were monitoring the affected services to confirm recovery. No further update followed, and the incident closed out at 1h 58m total. Two components were listed as affected: Web and ChatGPT. No user reports were filed during the window on this record, so the only account of impact is OpenAI's own two-update trail. That trail is thin: it names the symptom, errors on sending messages, and the plan tier, paid, but gives no scope figures and no cause. The 90 day history for OpenAI shows 100 incidents with a median duration of 1 hour 40 minutes, so this incident ran slightly longer than that median. Across all time OpenAI has logged 993 incidents, averaging 15.3 a month, which is useful context for how often this provider posts status updates of any size. OpenAI has not published a post-mortem for this incident, and nothing on the status page explains what caused the error rates in the first place.
Learning
Over the last 30 days, external monitoring of OpenAI recorded 42 incidents, all resolved, with a median duration of 1h 53m, a figure close to what this single incident ran. That overlap matters: it means an outage like this one, confined to paid ChatGPT plans sending messages, can run its full course while a basic uptime check on your own product still reads green, because your product's front door was never the thing that broke. The failure here sat inside OpenAI's own request handling, a layer your monitoring cannot see because it isn't yours to probe. If you only watch your own endpoints, an OpenAI-side error spike like this one shows up to you as a vague rise in failed requests with no clear source, and you lose time chasing your own stack for a fault that was never there. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.