| Benchmark | Gemini 2.5 Pro | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 7.2 | 30.9 |
| AA Intelligence | 27 | 46.7 |
| AA-LCR | 66 | 74.7 |
| AA-Omniscience | -14.3 | 13.5 |
| arena_elo | 1446 | 1474 |
| arena_text_factuality | 1462 | 1479 |
| Artificial Analysis Coding Index | 46.7 | 66 |
| coding_arena_elo | 1224 | 1517 |
| critpt | 2.6 | 13.4 |
| gdpval | 8.5 | 38.5 |
| GPQA Diamond | 84.4 | 92.4 |
| HLE | 22.5 | 41.4 |
| HLE (with tools) | 28.4 | 53.5 |
| hmmt_nov_2025 | 80 | 95 |
| IFBench | 49 | 80.5 |
| MMLU-Pro | 86 | 89.6 |
| OmniScience Accuracy | 39 | 31.1 |
| OmniScience Non-Hallucination | 12.6 | 74.4 |
| scicode | 43 | 53.5 |
| simplebench | 62.4 | 70.4 |
| simpleqa_verified | 56 | 58.5 |
| SWE-bench Verified | 67.2 | 80.4 |
| TauBench V3 - Banking | 9.7 | 11.8 |
| Terminal-Bench 2.1 | 28.5 | 75 |
| Terminal-Bench Hard | 26.5 | 50.8 |
| τ²-Bench Telecom (AA run) | 54.1 | 94.7 |
| τ³-Bench | 9.3 | 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.