| Benchmark | Claude 4 Opus | Gemini 3.5 Flash |
|---|---|---|
| AA Intelligence | 31.7 | 52 |
| AA-LCR | 40 | 81 |
| arc_agi_1 | 27 | 92.5 |
| ARC-AGI-2 | 8.6 | 72.1 |
| Artificial Analysis Coding Index | 34 | 70.1 |
| frontiermath_tier_4 | 4.2 | 14.6 |
| GPQA Diamond | 79.6 | 92.2 |
| HLE | 12.3 | 42.7 |
| IFBench | 53.7 | 76.3 |
| livebench | 74.6 | 75 |
| scicode | 40.9 | 53.1 |
| simplebench | 58.8 | 76.7 |
| Terminal-Bench Hard | 31.1 | 46.2 |
| τ²-Bench Telecom (AA run) | 73.4 | 95.6 |
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.