| Benchmark | Claude Sonnet 5 | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 49.7 | 30.9 |
| AA Intelligence | 55.3 | 46.7 |
| AA-LCR | 77 | 74.7 |
| AA-Omniscience | 16.4 | 13.5 |
| arena_elo | 1463 | 1474 |
| arena_text_factuality | 1453 | 1479 |
| Artificial Analysis Coding Index | 71.5 | 66 |
| coding_arena_elo | 1540 | 1517 |
| critpt | 16.9 | 13.4 |
| DeepSWE 1.1 | 54 | 21.6 |
| gdpval | 54.8 | 38.5 |
| GPQA Diamond | 91.1 | 92.4 |
| HLE | 57.4 | 41.4 |
| livebench_agentic_coding | 59.4 | 43.6 |
| livebench_coding | 80.7 | 74.2 |
| livebench_data_analysis | 71.7 | 71.8 |
| livebench_instruction_following | 63.9 | 74 |
| livebench_language | 75 | 79.7 |
| livebench_math | 92.9 | 85.3 |
| livebench_reasoning | 88.7 | 83.3 |
| OmniScience Accuracy | 40 | 31.1 |
| OmniScience Non-Hallucination | 60.6 | 74.4 |
| scicode | 53.6 | 53.5 |
| simplebench | 60.6 | 70.4 |
| simpleqa_verified | 25 | 58.5 |
| SWE-bench Multilingual | 78.3 | 78.3 |
| SWE-bench Pro | 63.2 | 60.6 |
| SWE-bench Verified | 85.2 | 80.4 |
| TauBench V3 - Banking | 37.3 | 11.8 |
| Terminal-Bench 2.0 | 80.4 | 69.7 |
| Terminal-Bench 2.1 | 80.5 | 75 |
| τ³-Bench | 28.2 | 10.9 |
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.