BrokenAI

An independent reliability ledger for the major model providers. Built because their status pages weren't enough.

Pages
ScoreboardAbout & scoringSLA compliance30-day report
Providers
ClaudeOpenAIMistralGrokDeepSeekKimiCohere
Integrate
Status badgellms.txtSitemaprobots.txt
© 2026 BrokenAI · Independent · Not affiliated with any provider · See a number that looks wrong? @threatner_v2.1.0
BrokenAI
30-day reportSLAEmbedAbout
30-day reliability report

How did the major AI providers behave this month?

Numbers come straight from each provider's official Statuspage, recomputed on every visit. The greener cell on a row is the provider that came out ahead; ties get no tint. Hover a row for the metric's definition.

Metric
Claude
OpenAI
Mistral
Grok
DeepSeek
Kimi
Cohere
  1. Reliability
    1. Composite reliability score
      50.7
      58.8
      70.0
      88.3
      87.6
      58.3
      ▲89.8
    2. Uptime
      98.677%
      99.643%
      99.915%
      100.000%
      99.922%
      98.869%
      100.000%
    3. Total downtime
      9h 31m
      2h 34m
      37m
      0m
      34m
      8h 8m
      0m
    4. Active right now
      0
      0
      0
      0
      0
      0
      0
  2. Incidents
    1. Total incidents
      46
      36
      131
      5
      1
      40
      1
    2. Major + critical
      18
      4
      1
      0
      1
      40
      0
    3. Minor
      27
      28
      130
      5
      0
      0
      0
    4. Longest single incident
      5h 44m
      32h 4m
      29h 33m
      10h 2m
      ▲34m
      1h 35m
      2h 54m
    5. Longest clean streak
      3d 15hMay 22–26
      6dMay 13–19
      3d 5hJun 3–7
      8d 3hMay 18–26
      29d 14hMay 8 – Jun 7
      8d 1hMay 14–22
      ▲
      30dMay 8 – Jun 7
  3. Time to fix
    1. Median fix time
      39m
      1h 31m
      5m
      1h 44m
      34m
      ▲3m
      2h 54m
    2. On a bad day
      3h 9m
      13h 2m
      ▲19m
      10h 2m
      34m
      37m
      2h 54m
    3. Closed vs still open
      46 closedall wrapped up
      35 closedall wrapped up
      131 closedall wrapped up
      5 closedall wrapped up
      1 closedall wrapped up
      40 closedall wrapped up
      1 closedall wrapped up
  4. Patterns
    1. What broke most
      Claude Code39 incidents · 18h 47m attributed
      Conversations13 incidents · 32h 7m attributed
      Conversations API35 incidents · 5h 53m attributed
      API Console2 incidents · 11h 46m attributed
      API 服务 (API Service)1 incident · 34m attributed
      Agentic Model23 incidents · 8h 11m attributed
      Unspecified1 incident · 2h 54m attributed
    2. Most common failure type
      Elevated errors38 of 46
      Elevated errors13 of 36
      Degraded95 of 131
      Degraded2 of 5
      Other1 of 1
      Other40 of 40
      Other1 of 1
    3. Busiest day of week
      Friday11 incidents on this day
      Friday9 incidents on this day
      Monday36 incidents on this day
      Wednesday2 incidents on this day
      Too few to call
      Friday15 incidents on this day
      Too few to call
    4. Busiest hour of day
      8 AM UTC5 incidents started here
      5 PM UTC5 incidents started here
      4 PM UTC17 incidents started here
      3 PM UTC2 incidents started here
      Too few to call
      9 AM UTC8 incidents started here
      Too few to call
What this report doesn't cover
  • 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.