| Benchmark | Muse Spark 1.1 | Qwen3.7 Plus Preview |
|---|---|---|
| AA Agentic Index | 39.7 | 20.8 |
| AA Intelligence | 53.2 | 39.4 |
| AA-LCR | 81.3 | 69 |
| AA-Omniscience | 28.1 | 2.4 |
| arena_elo | 1491 | 1458 |
| arena_text_factuality | 1466 | 1454 |
| Artificial Analysis Coding Index | 71.3 | 55.9 |
| baby_vision | 76.3 | 70.4 |
| critpt | 15.1 | 9.1 |
| DeepSWE 1.1 | 53.3 | 14.2 |
| Finance Agent v2 | 57.2 | 38.2 |
| gdpval | 43.7 | 22.2 |
| GPQA Diamond | 89.8 | 90.3 |
| HLE | 62.1 | 35.6 |
| JobBench | 54.7 | 27.6 |
| MCP Atlas | 88.1 | 73.2 |
| OmniScience Accuracy | 52 | 22.5 |
| OmniScience Non-Hallucination | 50 | 74.5 |
| OSWorld-Verified | 80.8 | 73.3 |
| scicode | 58.2 | 51.3 |
| SWE-bench Pro | 61.5 | 57.6 |
| TauBench V3 - Banking | 31.8 | 17.9 |
| Terminal-Bench 2.1 | 80 | 61 |
| WebArena-Verified | 69 | 55.3 |
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.