| Benchmark | Qwen3.6 27B | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 27.5 | 30.9 |
| AA Intelligence | 37.7 | 46.7 |
| AA-LCR | 73.3 | 74.7 |
| AA-Omniscience | -20 | 13.5 |
| aime_2026 | 94.1 | 97 |
| Artificial Analysis Coding Index | 53.7 | 66 |
| Claw Eval (pass@3) | 60.6 | 65.2 |
| critpt | 1.1 | 13.4 |
| DeepSWE 1.1 | 13.3 | 21.6 |
| gdpval | 32 | 38.5 |
| GPQA Diamond | 87.8 | 92.4 |
| HLE | 24 | 41.4 |
| hmmt_feb_2026 | 84.3 | 97.1 |
| hmmt_nov_2025 | 90.7 | 95 |
| IFBench | 69.1 | 80.5 |
| imo_answer_bench | 80.8 | 90 |
| JobBench | 21.8 | 31.3 |
| livebench_agentic_coding | 39.3 | 43.6 |
| livebench_coding | 71.8 | 74.2 |
| livebench_data_analysis | 70.4 | 71.8 |
| livebench_instruction_following | 53.2 | 74 |
| livebench_language | 63.3 | 79.7 |
| livebench_math | 79.9 | 85.3 |
| livebench_reasoning | 70.3 | 83.3 |
| LiveCodeBench v6 | 83.9 | 91.6 |
| mmlu_redux | 93.5 | 95 |
| MMLU-Pro | 86.2 | 89.6 |
| nl2repo | 36.2 | 47.2 |
| OmniScience Accuracy | 19.6 | 31.1 |
| OmniScience Non-Hallucination | 50.7 | 74.4 |
| scicode | 39.8 | 53.5 |
| supergpqa | 66 | 73.6 |
| SWE-bench Multilingual | 71.3 | 78.3 |
| SWE-bench Pro | 53.5 | 60.6 |
| SWE-bench Verified | 77.2 | 80.4 |
| TauBench V3 - Banking | 16.7 | 11.8 |
| Terminal-Bench 2.0 | 59.3 | 69.7 |
| Terminal-Bench 2.1 | 60.7 | 75 |
| Terminal-Bench Hard | 34.8 | 50.8 |
| τ²-Bench Telecom (AA run) | 94.2 | 94.7 |
| τ³-Bench | 15.3 | 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.