Historical SLA

AI API Uptime History & Incident Log

Uptime, outages, degradations and time-to-resolve, computed from each vendor's published incident history. Major/critical incidents count as downtime; minor ones as degradation.

OpenAI

100.000%

Incidents
1
Outages
0
Downtime
0 min
Degraded
1.3 h
Mean time to resolve
1.3 h
Unconfirmed (not counted)
3

Claude / Anthropic

97.113%

Incidents
7
Outages
7
Downtime
4.8 h
Degraded
0 min
Mean time to resolve
42 min
Unconfirmed (not counted)
0

Perplexity

100.000%

Incidents
0
Outages
0
Downtime
0 min
Degraded
0 min
Mean time to resolve
—
Unconfirmed (not counted)
0

OpenRouter

100.000%

Incidents
0
Outages
0
Downtime
0 min
Degraded
0 min
Mean time to resolve
—
Unconfirmed (not counted)
0

GitHub / Copilot

99.252%

Incidents
2
Outages
1
Downtime
1.3 h
Degraded
42 min
Mean time to resolve
59 min
Unconfirmed (not counted)
3

Cloudflare Workers AI

100.000%

Incidents
1
Outages
0
Downtime
0 min
Degraded
39 min
Mean time to resolve
39 min
Unconfirmed (not counted)
4
As of: Oct 29, 2026, 02:42 PM UTCSource: Providers' public status and incident feeds Method: Only incidents whose affected components serve AI APIs are counted (e.g. Cloudflare Workers AI/AI Gateway, OpenAI API services, Claude API, GitHub Copilot); incidents without component data are listed as platform-wide but not counted. Major/critical duration counts as downtime, minor as degraded — a derived observation, not a provider-certified SLA.

Incident audit log · last 7 days

Get paged when this happens — alerts can go to email, Slack, Discord, PagerDuty or Opsgenie.

Set up alerts
Help