| Benchmark | Gemma 4 E2B | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 7.4 | 39.7 |
| AA Intelligence | 10 | 53.2 |
| AA-LCR | 17 | 81.3 |
| AA-Omniscience | -23.6 | 28.1 |
| Artificial Analysis Coding Index | 8.3 | 71.3 |
| critpt | 0 | 15.1 |
| gdpval | 0 | 43.7 |
| GPQA Diamond | 43.4 | 89.8 |
| HLE | 4.8 | 62.1 |
| OmniScience Accuracy | 8.2 | 52 |
| OmniScience Non-Hallucination | 67.6 | 50 |
| scicode | 20.9 | 58.2 |
| TauBench V3 - Banking | 4.5 | 31.8 |
| Terminal-Bench 2.1 | 0.4 | 80 |
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.