Qwen vision-language model for visual reasoning, documents, and agent tasks
Benchmarks
| Benchmark | Category | Score | Variant | Provenance | Source | Observed |
|---|---|---|---|---|---|---|
| AA Agentic Index | agentic | 20.7 | — | mirror | openrouter | Jul 23, 2026 |
| 15.8 | non-reasoning | independent | aa | Jul 24, 2026 | ||
| 20.7 | — | independent | aa | Jul 24, 2026 | ||
| AA Coding Index | coding | 45.7 | — | mirror | openrouter | Jul 23, 2026 |
| 43.3 | non-reasoning | independent | aa | Jul 24, 2026 | ||
| 45.7 | — | independent | aa | Jul 24, 2026 | ||
| AA Intelligence Index | intelligence | 32.3 | — | mirror | openrouter | Jul 23, 2026 |
| 27.6 | non-reasoning | independent | aa | Jul 24, 2026 | ||
| 32.3 | — | independent | aa | Jul 24, 2026 | ||
| LMArena Text | preference | 1432 | — | crowd | lmarena | Jul 21, 2026 |
| LMArena Vision | preference | 1246 | — | crowd | lmarena | Jul 21, 2026 |
| LMArena WebDev | preference | 1364 | — | crowd | lmarena | Jul 21, 2026 |
| Median output speed | speed | 145 tok/s | non-reasoning | independent | aa | Jul 24, 2026 |
| 134 tok/s | — | independent | aa | Jul 24, 2026 | ||
| Median time to first token | speed | 2.36s | non-reasoning | independent | aa | Jul 24, 2026 |
| 2.37s | — | independent | aa | Jul 24, 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 | $0.26 · $2.08 | 82.33% | — | — |
| AtlasCloud | fp8 | 262K | $0.3 · $2.4 | 86.90% | — | — |
| DeepInfra | fp4 | 262K | $0.29 · $2.4 | 81.86% | — | — |
| Novita | bf16 | 262K | $0.4 · $3.2 | 93.56% | — | — |
| SiliconFlow | fp8 | 262K | $0.26 · $2.08 | 52.00% | — | — |