| Benchmark | GPT-5.6 Sol | Qwen3.7 Plus Preview |
|---|---|---|
| AA Agentic Index | 57.8 | 20.8 |
| AA Intelligence | 61 | 39.4 |
| AA-LCR | 77.7 | 69 |
| AA-Omniscience | 22 | 2.4 |
| Agents' Last Exam | 52.7 | 33.6 |
| arena_elo | 1482 | 1458 |
| arena_text_factuality | 1474 | 1454 |
| Artificial Analysis Coding Index | 78.3 | 55.9 |
| charxiv_rq | 87.8 | 85.9 |
| critpt | 32.3 | 9.1 |
| DeepSWE 1.1 | 73 | 14.2 |
| Finance Agent v2 | 53.8 | 38.2 |
| gdpval | 61.1 | 22.2 |
| GPQA Diamond | 94.6 | 90.3 |
| HLE | 49.5 | 35.6 |
| IFBench | 72.7 | 79.1 |
| JobBench | 45.4 | 27.6 |
| mathvision | 95.8 | 90.3 |
| MCP Atlas | 83.6 | 73.2 |
| MMMU-Pro | 84.6 | 80.5 |
| OmniScience Accuracy | 59.4 | 22.5 |
| OmniScience Non-Hallucination | 10.6 | 74.5 |
| OSWorld-Verified | 83 | 73.3 |
| scicode | 56.9 | 51.3 |
| SWE-bench Pro | 64.6 | 57.6 |
| TauBench V3 - Banking | 44.3 | 17.9 |
| Terminal-Bench 2.1 | 89.5 | 61 |
| Terminal-Bench Hard | 65.9 | 47 |
| Video-MME | 89.5 | 88 |
| τ²-Bench Telecom (AA run) | 85.1 | 93 |
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.