| Benchmark | Muse Spark | Qwen3.7 Plus Preview |
|---|---|---|
| AA Agentic Index | 28.7 | 20.8 |
| AA Intelligence | 44.3 | 39.4 |
| AA-LCR | 77 | 69 |
| AA-Omniscience | 7.2 | 2.4 |
| arena_elo | 1488 | 1458 |
| arena_text_factuality | 1467 | 1454 |
| arena_vision | 1295 | 1263 |
| Artificial Analysis Coding Index | 58.6 | 55.9 |
| critpt | 11.3 | 9.1 |
| ERQA | 64.7 | 69.8 |
| gdpval | 32.3 | 22.2 |
| GPQA Diamond | 89.5 | 90.3 |
| HLE | 58.4 | 35.6 |
| IFBench | 75.9 | 79.1 |
| MCP Atlas | 82.2 | 73.2 |
| MMMU-Pro | 80.5 | 80.5 |
| OmniScience Accuracy | 49.6 | 22.5 |
| OmniScience Non-Hallucination | 26.8 | 74.5 |
| scicode | 51.5 | 51.3 |
| screenspot_pro_no_tools | 84.1 | 79 |
| SimpleVQA | 0.7 | 0.8 |
| SWE-bench Pro | 55 | 57.6 |
| SWE-bench Verified | 77.4 | 77.7 |
| TauBench V3 - Banking | 19.6 | 17.9 |
| Terminal-Bench 2.0 | 59 | 70.3 |
| Terminal-Bench 2.1 | 62.2 | 61 |
| Terminal-Bench Hard | 45.5 | 47 |
| τ²-Bench Telecom (AA run) | 91.5 | 93 |
| τ³-Bench | 19.6 | 17.9 |
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.