| Benchmark | Gemini 3.5 Flash | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 70.4 | 30.9 |
| AA Intelligence | 52 | 46.7 |
| AA-LCR | 81 | 74.7 |
| AA-Omniscience | 21.2 | 13.5 |
| aime_2026 | 95 | 97 |
| arena_elo | 1477 | 1474 |
| arena_text_factuality | 1466 | 1479 |
| Artificial Analysis Coding Index | 70.1 | 66 |
| coding_arena_elo | 1490 | 1517 |
| critpt | 13.1 | 13.4 |
| DeepSWE | 37 | 18 |
| DeepSWE 1.1 | 37 | 21.6 |
| Finance Agent v2 | 57.9 | 48.4 |
| gdpval | 57.8 | 38.5 |
| GPQA Diamond | 92.2 | 92.4 |
| HLE | 42.7 | 41.4 |
| hmmt_feb_2026 | 95.5 | 97.1 |
| IFBench | 76.3 | 80.5 |
| itbenchSre | 40.3 | 42.5 |
| livebench | 75 | 74.3 |
| livebench_agentic_coding | 49 | 43.6 |
| livebench_coding | 78.2 | 74.2 |
| livebench_data_analysis | 64.9 | 71.8 |
| livebench_instruction_following | 75.6 | 74 |
| livebench_language | 84.6 | 79.7 |
| livebench_math | 88.2 | 85.3 |
| livebench_reasoning | 82 | 83.3 |
| MCP Atlas | 83.6 | 76.4 |
| OmniScience Accuracy | 51.4 | 31.1 |
| OmniScience Non-Hallucination | 38.2 | 74.4 |
| scicode | 53.1 | 53.5 |
| simplebench | 76.7 | 70.4 |
| simpleqa_verified | 68.4 | 58.5 |
| SWE-bench Pro | 55.1 | 60.6 |
| TauBench V3 - Banking | 32.2 | 11.8 |
| Terminal-Bench 2.0 | 76.2 | 69.7 |
| Terminal-Bench 2.1 | 78.7 | 75 |
| Terminal-Bench Hard | 46.2 | 50.8 |
| τ²-Bench Telecom (AA run) | 95.6 | 94.7 |
| τ³-Bench | 25.4 | 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.