Qwen vision-language model for visual reasoning, documents, and agent tasks
Benchmarks
| Benchmark | Category | Score | Variant | Provenance | Source | Observed |
|---|---|---|---|---|---|---|
| AA Agentic Index | agentic | 11.8 | — | mirror | openrouter | Jul 23, 2026 |
| AA Coding Index | coding | 37.0 | — | mirror | openrouter | Jul 23, 2026 |
| AA Intelligence Index | intelligence | 24.3 | — | mirror | openrouter | Aug 7, 2026 |
| 19.3 | — | independent | aa | Sep 9, 2026 | ||
| LMArena Text | preference | 1395 | — | crowd | lmarena | Sep 13, 2026 |
| LMArena WebDev | preference | 1250 | — | crowd | lmarena | Sep 11, 2026 |
| Median output speed | speed | 147 tok/s | — | independent | aa | Sep 21, 2026 |
| Median time to first token | speed | 2.09s | — | independent | aa | Sep 21, 2026 |
Includes index data from Artificial Analysis.
Providers
Live operational stats per hosting endpoint, via OpenRouter.
| Provider | Quant | Context | $/1M in · out | Uptime 24h | Latency | Throughput |
|---|---|---|---|---|---|---|
| Alibaba | — | 262K | $0.163 · $1.3 | 98.50% | — | — |
| AtlasCloud | fp8 | 262K | $0.225 · $1.8 | 99.39% | — | — |
| Darkbloom | fp4 | 262K | $0.08 · $0.75 | 99.93% | — | — |
| DeepInfra | fp8 | 262K | $0.14 · $1 | 99.96% | — | — |
| Parasail | fp8 | 262K | $0.15 · $1 | 99.77% | — | — |
| SiliconFlow | fp8 | 262K | $0.24 · $1.8 | 67.16% | — | — |
| Venice | — | 256K | $0.313 · $1.25 | 99.97% | — | — |