| Benchmark | GLM-5.1 | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 66 | 30.9 |
| AA Intelligence | 41 | 46.7 |
| AA-LCR | 68 | 74.7 |
| AA-Omniscience | 0.8 | 13.5 |
| aime_2026 | 95.8 | 97 |
| arena_elo | 1468 | 1474 |
| arena_text_factuality | 1458 | 1479 |
| Artificial Analysis Coding Index | 55.8 | 66 |
| coding_arena_elo | 1510 | 1517 |
| critpt | 4.6 | 13.4 |
| DeepSWE | 18 | 18 |
| Finance Agent v2 | 44.8 | 48.4 |
| gdpval | 49.5 | 38.5 |
| GPQA Diamond | 86.8 | 92.4 |
| HLE | 52.3 | 41.4 |
| HLE (with tools) | 52.3 | 53.5 |
| hmmt_feb_2026 | 89.4 | 97.1 |
| hmmt_nov_2025 | 94 | 95 |
| IFBench | 76.3 | 80.5 |
| imo_answer_bench | 83.8 | 90 |
| itbenchSre | 40.3 | 42.5 |
| livebench | 70.2 | 74.3 |
| MCP Atlas | 75.6 | 76.4 |
| MMLU-Pro | 86 | 89.6 |
| nl2repo | 42.7 | 47.2 |
| OmniScience Accuracy | 25.2 | 31.1 |
| OmniScience Non-Hallucination | 70.1 | 74.4 |
| scicode | 43.8 | 53.5 |
| simplebench | 55.1 | 70.4 |
| simpleqa_verified | 38.1 | 58.5 |
| SWE-bench Multilingual | 73.3 | 78.3 |
| SWE-bench Pro | 58.4 | 60.6 |
| TauBench V3 - Banking | 13.6 | 11.8 |
| Terminal-Bench 2.0 | 69 | 69.7 |
| Terminal-Bench 2.1 | 63.5 | 75 |
| Terminal-Bench Hard | 43.2 | 50.8 |
| τ²-Bench Telecom (AA run) | 97.7 | 94.7 |
| τ³-Bench | 70.6 | 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.