OpenAI API outage on October 21, 2026: Elevated error rates for gpt-5.6-luna API requests
OpenAI reported an incident affecting OpenAI API for 1 hour 9 minutes on 21 October 2026, from 15:34 to 16:44 UTC.
Resolved
Official incident reported by OpenAI
- Started
- 15:34
- 21 Oct 2026 UTC
- Resolved
- 16:44
- 21 Oct 2026 UTC
- Duration
- 1h 9m
- 1 hour 9 minutes
- Impact
- Minor
- as rated by OpenAI
- Updates
- 3
- posted by OpenAI
Timeline
All times UTC.
Official incident from OpenAI
Elevated error rates for gpt-5.6-luna API requests
-
Official
Identified
We have identified that users are experiencing elevated errors for the impacted services. We are working on implementing a mitigation.
-
Official
Monitoring
We have applied the mitigation and are monitoring the recovery.
-
Official
Resolved
All impacted services have now fully recovered.
"Official" rows are quoted from OpenAI's public status page, status.openai.com. Very long posts are shortened.
What our checkpoints saw
HostTracker's checkpoints, which test api.openai.com from several locations, recorded no outage of 5 minutes or longer during this incident. A provider can report a problem that affects only some features, regions or accounts, which a check of the main address does not see. The live OpenAI API status shows what our checks see right now.
Does your product depend on OpenAI API?
An outage like this one reaches your users through your product, and most teams hear about it from a customer rather than from a dashboard. HostTracker watches any endpoint you name - your own site, an API you call, a vendor you resell - from 300+ locations, confirms a failure from several of them before it says anything, and then tells you by email, SMS, voice call or your chat tool.
You get the same record you are reading here for your own services: every outage timestamped, with the checks that observed it.
Independent measurement by HostTracker. Not affiliated with OpenAI API. Official updates are quoted from OpenAI's public status page and remain OpenAI's own words.