The AI-provider API latency and per-feed uptime we measure ourselves on a fixed schedule — vendor-neutral, time-stamped, and kept. A backfill-proof archive that can't be reconstructed after the fact.
| Provider | Avg latency | Uptime | Samples |
|---|---|---|---|
| Groq | 63 ms | 100.0% | 144 |
| xAI | 67 ms | 100.0% | 144 |
| OpenAI | 70 ms | 99.3% | 144 |
| 77 ms | 100.0% | 144 | |
| Anthropic | 107 ms | 100.0% | 144 |
| Mistral | 191 ms | 100.0% | 144 |
| DeepSeek | 317 ms | 100.0% | 144 |
Live JSON: /benchmarks · embeddable widget: /widget/latency
overall measured uptime across all live tools, over the selected window.
Live status board: /uptime · status JSON: /status.json