Data sources: Status is fetched directly from each provider's official public status feed
(Atlassian Statuspage JSON API) where available.
Providers marked manual check don't expose a public machine-readable status API , links go to their
status pages.
Response time reflects the latency measured fetching the status endpoint , not AI inference
latency.
Typical latency shows average Time-to-First-Token benchmarks from public sources (Artificial
Analysis, etc.), updated quarterly.
90-day history is generated from provider-specific uptime reports where available.