| Benchmark | Muse Spark | Qwen3.6 27B |
|---|---|---|
| AA Agentic Index | 28.7 | 27.5 |
| AA Intelligence | 44.3 | 37.7 |
| AA-LCR | 77 | 73.3 |
| AA-Omniscience | 7.2 | -20 |
| Artificial Analysis Coding Index | 58.6 | 53.7 |
| critpt | 11.3 | 1.1 |
| ERQA | 64.7 | 62.5 |
| gdpval | 32.3 | 32 |
| GPQA Diamond | 89.5 | 87.8 |
| HLE | 58.4 | 24 |
| IFBench | 75.9 | 69.1 |
| MMMU-Pro | 80.5 | 75.8 |
| OmniScience Accuracy | 49.6 | 19.6 |
| OmniScience Non-Hallucination | 26.8 | 50.7 |
| scicode | 51.5 | 39.8 |
| SimpleVQA | 0.7 | 56.1 |
| SWE-bench Pro | 55 | 53.5 |
| SWE-bench Verified | 77.4 | 77.2 |
| TauBench V3 - Banking | 19.6 | 16.7 |
| Terminal-Bench 2.0 | 59 | 59.3 |
| Terminal-Bench 2.1 | 62.2 | 60.7 |
| Terminal-Bench Hard | 45.5 | 34.8 |
| τ²-Bench Telecom (AA run) | 91.5 | 94.2 |
| τ³-Bench | 19.6 | 15.3 |
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.