The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...
Share of successful requests per carrier, measured over the last 5 minutes, 30 minutes and day. When one carrier dips, the router fails over to the next.
| Provider | Status | Last 5m | Last 30m | Last day |
|---|---|---|---|---|
| Operational | 100.00% | 100.00% | 99.77% | |
| Operational | 100.00% | 100.00% | 99.18% | |
| Operational | — | 100.00% | 100.00% | |
| Operational | — | — | 79.05% | |
| Operational | 100.00% | 100.00% | 99.48% | |
| Operational | — | 100.00% | 90.26% | |
| Operational | — | — | 99.79% | |
| Operational | — | 0.00% | 75.25% | |
| Operational | — | — | 99.83% | |
| Operational | 100.00% | 100.00% | 99.61% |