If you pay OpenAI with a U.S. business Mastercard, you likely saw payment declines starting 18:38 UTC on 6 November, lasting 93h 29m. OpenAI says the fault sat with its payment partners, not with your account or integration.
Started
18:38 UTC
Duration
Lasted 3d 21h
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.
What happened?
OpenAI's status page opened this incident at 18:38 UTC, reporting payment declines on specific U.S. business Mastercard ranges: cards beginning 544015, 546400, 543502, 543799, 512998, and some beginning 513989. The first update was explicit about the knock-on effect for API customers: automatic recharges and one-time credit purchases could fail, and if your prepaid balance ran out, your API requests would fail too. OpenAI's own advice was to add a card from a different issuer before your next payment, even without seeing a decline yourself, and to retry any failed credit purchase on the new card. No further detail was given on how many accounts were affected, since OpenAI did not publish a count. The next and only other update came at 01:29 UTC, saying a mitigation had been applied and the services were being monitored to confirm recovery. That update did not explain what the mitigation was or why it took until then to apply. The incident ran for 93h 29m in total before being marked resolved, with no updates in between confirming progress. No components list was published for this incident, and no user reports were logged in the window. For a payments fault of this length, two updates is thin. You're left relying on OpenAI's own account of what happened, since there is nothing else on the record to check it against.
Learning
This was a billing and payments failure, not an API or model outage, so uptime checks against OpenAI's endpoints would have stayed green throughout, correctly, because the service itself kept responding. The break sat one layer up, in the card processing path between OpenAI and its payment partners, somewhere your own monitoring has no visibility into at all. If your API requests started failing partway through this window, it was because your prepaid balance ran dry while recharges were silently declining, not because anything in your integration broke. OpenAI's 90-day history shows 101 incidents with a median duration of 1 hour 50 minutes, so an incident running over 93 hours is far outside its normal pattern, yet the gap between the two updates gives you no way to judge what was happening in between. The general shape to recognise: a payment or billing fault can look identical from your side to a quota or rate-limit problem, both show up as failed requests, and only the provider's own status page tells them apart. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.