Cohere Outage History

9 incidents recorded over the last 365 days, 9 of them resolved. Durations are measured from when PulsAPI first saw the problem to when it cleared, which is usually longer than the vendor's own figure.

Every Recorded Cohere Incident

Cohere incidents, newest first, with date, duration and status.
DateIncidentDurationStatus
Aug 5, 2026Documentation Content Loading IssuesšŸŽ‰ The issue has been resolved and we are back to normal.57 minResolved
Jul 23, 2026Dashboard Login Service DisruptionšŸŽ‰ The issue affecting Dashboard login has been resolved. All Cohere services are operating normally. Thank you for your patience while we worked through this.50 minResolved
Jun 1, 2026/audio/transcriptions endpoint returns error for audio files with unexpected sampling ratešŸŽ‰ The issue has been resolved and we are back to normal.2h 54mResolved
Apr 20, 2026Docs Platform Degraded PerformanceDocs are back up!12 minResolved
Apr 13, 2026Multiple Model DisruptionsšŸŽ‰ The issue has been resolved and we are back to normal.1h 24mResolved
Apr 11, 2026Scheduled DB Maintenance: Model authentication and API access momentarily affectedšŸŽ‰ The issue has been resolved and we are back to normal.8 minResolved
Feb 13, 2026Some chat models are unresponsive or timing outThe issue has been resolved and all models are back to normal operation1h 17mResolved
Oct 18, 2025Chat models return 503 service error when /v1/chat requests omit message and chat_history while providing tool_resultsšŸŽ‰ The issue has been resolved and we are back to normal.36 minResolved
Oct 1, 2025logprobs param degeraded for `command-a-03-2025` and `command-a-reasoning-08-2025``logprobs` feature has been correctly marked as unavailable for the specified models.3h 49mResolved

How to Read This Cohere Incident Log

Each row is an incident PulsAPI observed, not a summary written afterwards. The duration is wall-clock time between the first failing check and the first clean one, so it includes the window before Cohere acknowledged anything. Vendor post-mortems typically measure from acknowledgement, which is why their numbers are usually shorter.

Incidents still open have no duration yet and are listed as ongoing rather than being given a running total. The archive covers the last 365 days; anything older has aged out of the window rather than never having happened.

The longest single Cohere outage in this window ran 3h 49m, against a mean recovery of 1h 21m. If you depend on Cohere in a customer-facing path, the longest figure is the one to design around. The mean is what happens on a normal bad day; the maximum is what happens on the worst one.