| Benchmark | Claude Opus 4.7 | Qwen3.6 35B A3B |
|---|---|---|
| AA Agentic Index | 64.6 | 21.6 |
| AA Intelligence | 55 | 32.1 |
| AA-LCR | 75.3 | 66.7 |
| AA-Omniscience | 27.3 | -22.2 |
| aime_2026 | 95.8 | 92.7 |
| Artificial Analysis Coding Index | 73.6 | 41.9 |
| browsecomp | 79.8 | 48.7 |
| charxiv_rq | 82.1 | 78 |
| critpt | 12 | 0.3 |
| gdpval | 58.6 | 27.8 |
| GPQA Diamond | 94.2 | 86 |
| HLE | 54.7 | 22.2 |
| hmmt_feb_2026 | 93.9 | 83.6 |
| IFBench | 58.6 | 64.4 |
| MCP Atlas | 79.1 | 62.8 |
| MMMU-Pro | 78.8 | 75.3 |
| OmniScience Accuracy | 48.9 | 18.8 |
| OmniScience Non-Hallucination | 57.7 | 49.5 |
| scicode | 54.5 | 35.8 |
| SWE-bench Multilingual | 80.5 | 67.2 |
| SWE-bench Pro | 64.3 | 49.5 |
| SWE-bench Verified | 87.6 | 73.4 |
| TauBench V3 - Banking | 34.6 | 9.3 |
| Terminal-Bench 2.0 | 69.4 | 51.5 |
| Terminal-Bench 2.1 | 83.1 | 44.9 |
| Terminal-Bench Hard | 54.5 | 34.8 |
| toolathlon | 59.3 | 26.9 |
| τ²-Bench Telecom (AA run) | 88.6 | 95.3 |
| τ³-Bench | 28.9 | 67.2 |
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.