| Benchmark | Claude 4 Opus | GPT-5.6 Sol |
|---|---|---|
| AA Intelligence | 31.7 | 61 |
| AA-LCR | 40 | 77.7 |
| arc_agi_1 | 27 | 97.5 |
| ARC-AGI-2 | 8.6 | 92.5 |
| Artificial Analysis Coding Index | 34 | 78.3 |
| GPQA Diamond | 79.6 | 94.6 |
| HLE | 12.3 | 49.5 |
| IFBench | 53.7 | 72.7 |
| longbench_v2 | 55.6 | 67.1 |
| scicode | 40.9 | 56.9 |
| simplebench | 58.8 | 64.8 |
| Terminal-Bench Hard | 31.1 | 65.9 |
| τ²-Bench Telecom (AA run) | 73.4 | 85.1 |
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.