AI Infrastructure Reliability
Which AI API is most reliable?
Compare uptime scores, incident frequency, and mean resolution times across every major LLM provider. Updated daily from real incident data.
AI API Providers
Google Cloud
100Excellent · 100.000% uptime
0
Incidents
0
Last 90d
100.000%
Uptime
Mistral AI
100Excellent · 100.000% uptime
0
Incidents
0
Last 90d
100.000%
Uptime
Together AI
100Excellent · 100.000% uptime
0
Incidents
0
Last 90d
100.000%
Uptime
HuggingFace
100Excellent · 100.000% uptime
0
Incidents
0
Last 90d
100.000%
Uptime
Cohere
58.2Poor · 100.000% uptime
30
Incidents
4
Last 90d
100.000%
Uptime
Replicate
49.8Poor · 100.000% uptime
39
Incidents
12
Last 90d
100.000%
Uptime
OpenAI
38Poor · 100.000% uptime
131
Incidents
62
Last 90d
100.000%
Uptime
Anthropic
3.8Poor · 71.857% uptime
196
Incidents
97
Last 90d
71.857%
Uptime
Cloud Infrastructure
Full rankings →Stripe
100Excellent · 100.000% uptime · 0 incidents
Groq
100Excellent · 100.000% uptime · 0 incidents
AI21 Labs
96.7Excellent · 111.847% uptime · 3 incidents
Stability AI
92.8Good · 100.000% uptime · 1 incidents
Atlassian
67Poor · 98.898% uptime · 34 incidents
ElevenLabs
44.7Poor · 100.000% uptime · 54 incidents
How we calculate reliability scores
Data sources
We crawl official status pages every 5 minutes and track every incident from initial detection to resolution. Scores are based on real incident data, not synthetic uptime checks.
Scoring formula
The reliability score (0-100) factors in incident frequency, severity distribution, mean resolution time, and 90-day trend direction. Higher scores indicate fewer and shorter incidents.
Get notified when AI providers go down
Set up alerts for OpenAI, Anthropic, Google AI, and more. Delivered via Slack, webhook, or Google Chat.
Set up alerts