| Benchmark | Qwen3.7 Plus Preview | Qwen3 VL 4B (Reasoning) |
|---|---|---|
| AA Agentic Index | 20.8 | 14.4 |
| AA Intelligence | 39.4 | 7.7 |
| AA-LCR | 69 | 23 |
| AA-Omniscience | 2.4 | -68.4 |
| Artificial Analysis Coding Index | 55.9 | 6.7 |
| critpt | 9.1 | 0 |
| ERQA | 69.8 | 47.3 |
| gdpval | 22.2 | 13.8 |
| GPQA Diamond | 90.3 | 64.1 |
| HLE | 35.6 | 4.6 |
| IFBench | 79.1 | 36.6 |
| ifeval | 94.6 | 82.6 |
| include | 83 | 64.6 |
| LiveCodeBench v6 | 89.6 | 51.3 |
| LVBench | 76.2 | 53.5 |
| mathvision | 90.3 | 60 |
| mmlu_prox | 85.4 | 65 |
| mmlu_redux | 94.5 | 86 |
| MMLU-Pro | 88.5 | 73.6 |
| MMMU-Pro | 80.5 | 57 |
| OmniScience Accuracy | 22.5 | 12.2 |
| OmniScience Non-Hallucination | 74.5 | 8.2 |
| OSWorld-Verified | 73.3 | 31.4 |
| RealWorldQA | 86.9 | 73.2 |
| scicode | 51.3 | 17.1 |
| screenspot_pro_no_tools | 79 | 49.2 |
| supergpqa | 71.4 | 46.8 |
| Terminal-Bench Hard | 47 | 1.5 |
| Video-MME | 88 | 59.7 |
| VideoMMMU | 85.4 | 69.4 |
| τ²-Bench Telecom (AA run) | 93 | 15.5 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.