| Benchmark | Muse Spark 1.1 | Qwen3 VL 4B (Reasoning) |
|---|---|---|
| AA Agentic Index | 39.7 | 14.4 |
| AA Intelligence | 53.2 | 7.7 |
| AA-LCR | 81.3 | 23 |
| AA-Omniscience | 28.1 | -68.4 |
| Artificial Analysis Coding Index | 71.3 | 6.7 |
| critpt | 15.1 | 0 |
| gdpval | 43.7 | 13.8 |
| GPQA Diamond | 89.8 | 64.1 |
| HLE | 62.1 | 4.6 |
| OmniScience Accuracy | 52 | 12.2 |
| OmniScience Non-Hallucination | 50 | 8.2 |
| OSWorld-Verified | 80.8 | 31.4 |
| scicode | 58.2 | 17.1 |
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.