KINGY.AI / SIGNAL OFFICIAL SOURCES ONLY
THE RELIABILITY REPORT
Trust is a track record.
Monthly incident-free time and quarterly API comparisons, with explicit coverage and a reproducible method.
WHICH API WAS MOST RELIABLE?
The data has to earn that answer.
A six-provider winner is withheld until every provider has at least 95% collection coverage for the same UTC quarter and no unresolved scope or timing gaps.
| API provider | Incident-free time | Coverage | Reported duration | Mean recovery | Evidence |
|---|---|---|---|---|---|
| OpenAI | — | 0.28% | 57h 35m | 11h 45m | Scope / timing gaps |
| Anthropic | — | 0.28% | 150h 29m | 37h 37m | Scope / timing gaps |
| — | 0.28% | 0m | — | Collecting evidence | |
| xAI | — | 0.28% | 1m | 1m | Collecting evidence |
| Mistral | — | 0.28% | 0m | — | Scope / timing gaps |
| DeepSeek | — | 0.28% | 9h 10m | 1h 32m | Scope / timing gaps |
Comparison uses explicitly tagged API incidents. Google means Vertex AI. An incident affecting one model, feature or region counts as a reported incident interval; this is not a request-weighted availability measure or an SLA comparison. Full methodology ↗
Uptime by month
REPORTED INCIDENT-FREE TIME| Month (UTC) | Incident-free time | Collection coverage | Incidents | Mean recovery |
|---|---|---|---|---|
| Nov 2026 | Insufficient coverage | 2.28% | 2 | 3h 24m |
| Oct 2026 | Insufficient coverage | 0% | 23 | 4h 16m |
| Sep 2026 | Insufficient coverage | 0% | 0 | — |
| Aug 2026 | Insufficient coverage | 0% | 0 | — |
| Jul 2026 | Insufficient coverage | 0% | 0 | — |
| Jun 2026 | Insufficient coverage | 0% | 0 | — |
Incident counts reflect retained reports, not a complete historical census. Recovery averages resolved records with known timestamps; reporting windows may differ from actual impact. Downtime intervals are merged to avoid double counting.