| Benchmark | Claude 4 Opus | Gemma 4 E2B |
|---|---|---|
| AA Intelligence | 31.7 | 10 |
| AA-LCR | 40 | 17 |
| Artificial Analysis Coding Index | 34 | 8.3 |
| GPQA Diamond | 79.6 | 43.4 |
| HLE | 12.3 | 4.8 |
| IFBench | 53.7 | 38 |
| ifeval | 87.4 | 82.9 |
| LiveCodeBench v6 | 47.4 | 44 |
| MMLU-Pro | 86.6 | 60 |
| mmmlu | 88.8 | 67.4 |
| scicode | 40.9 | 20.9 |
| Tau2 retail | 81.8 | 18.9 |
| Terminal-Bench Hard | 31.1 | 3 |
| τ²-Bench | 57 | 24.5 |
| τ²-Bench Telecom (AA run) | 73.4 | 22.2 |
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.