| Benchmark | Muse Spark 1.1 | Qwen3.6 Plus |
|---|---|---|
| AA Agentic Index | 39.7 | 29 |
| AA Intelligence | 53.2 | 40.5 |
| AA-LCR | 81.3 | 72.3 |
| AA-Omniscience | 28.1 | 2.6 |
| arena_elo | 1491 | 1444 |
| arena_text_factuality | 1466 | 1438 |
| Artificial Analysis Coding Index | 71.3 | 54.5 |
| coding_arena_elo | 1539 | 1460 |
| critpt | 15.1 | 2.9 |
| Finance Agent v2 | 57.2 | 40.9 |
| gdpval | 43.7 | 32 |
| GPQA Diamond | 89.8 | 90.4 |
| HLE | 62.1 | 28.8 |
| livebench_agentic_coding | 58.5 | 41.4 |
| livebench_coding | 77.2 | 78.2 |
| livebench_data_analysis | 72.5 | 69.9 |
| livebench_instruction_following | 69.6 | 58.3 |
| livebench_language | 74.3 | 75 |
| livebench_math | 87.1 | 83.7 |
| livebench_reasoning | 87.7 | 75.8 |
| MCP Atlas | 88.1 | 74.1 |
| OmniScience Accuracy | 52 | 26.4 |
| OmniScience Non-Hallucination | 50 | 68 |
| OSWorld-Verified | 80.8 | 62.5 |
| scicode | 58.2 | 40.7 |
| SWE-bench Pro | 61.5 | 56.6 |
| TauBench V3 - Banking | 31.8 | 20.8 |
| Terminal-Bench 2.1 | 80 | 61.4 |
| toolathlon | 75.6 | 39.8 |
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.