Metric
Claude
OpenAI
Mistral
Grok
DeepSeek
Kimi
Cohere
- Reliability
- Composite reliability score36.353.6100.097.1100.085.5100.0
- Uptime92.467%99.340%100.000%100.000%100.000%99.868%100.000%
- Total downtime54h 14m4h 45m0m0m0m57m0m
- Active right now1000000
- Composite reliability score
- Incidents
- Total incidents114603090
- Major + critical8700090
- Minor13003000
- Longest single incident134h 12m70h 39m—28m—▲27m—
- Longest clean streak7d 11hSep 22–294d 12hSep 17–2230dSep 13 – Oct 1314d 17hSep 22 – Oct 630dSep 13 – Oct 139d 12hSep 19–2830dSep 13 – Oct 13
- Total incidents
- Time to fix
- Median fix time1h 13m1h 5m—26m—▲2m—
- On a bad day6h 16m9h 23m—28m—▲27m—
- Closed vs still open10 closed1 still open46 closedall wrapped up0 closedall wrapped up3 closedall wrapped up0 closedall wrapped up9 closedall wrapped up0 closedall wrapped up
- Median fix time
- Patterns
- What broke mostClaude Cowork6 incidents · 100h 47m attributedConversations17 incidents · 85h 44m attributed—Grok (Web)2 incidents · 32m attributed—Search7 incidents · 29m attributed—
- Most common failure typeElevated errors7 of 11Elevated errors22 of 46—Other2 of 3—Other9 of 9—
- Busiest day of weekTuesday4 incidents on this dayTuesday11 incidents on this dayToo few to callTuesday2 incidents on this dayToo few to callFriday4 incidents on this dayToo few to call
- Busiest hour of day1 PM UTC2 incidents started here4 PM UTC5 incidents started hereToo few to call12 AM UTC1 incident started hereToo few to call10 PM UTC3 incidents started hereToo few to call
- What broke most
- No synthetic API probing — we only see what each provider publishes on their Statuspage. Quietly degraded responses that never get acknowledged on the status page are invisible here.
- No regional breakdown — Statuspage feeds don't reliably expose region-level impact, so a US-only outage and a global one count the same.
- No request-volume weighting — “1 incident” means one acknowledged incident, regardless of how many calls it actually affected.