| Benchmark | GPT-5.6 Sol | Nemotron 3 Super 120B A12B |
|---|---|---|
| AA Agentic Index | 57.8 | 8.8 |
| AA Intelligence | 61 | 25.7 |
| AA-LCR | 77.7 | 60.3 |
| AA-Omniscience | 22 | -41.5 |
| aime_2026 | 99.9 | 91.7 |
| Artificial Analysis Coding Index | 78.3 | 37.7 |
| browsecomp | 90.4 | 31.3 |
| critpt | 32.3 | 3.1 |
| gdpval | 61.1 | 9.9 |
| GPQA Diamond | 94.6 | 82.7 |
| HLE | 49.5 | 22.8 |
| IFBench | 72.7 | 72.6 |
| itbenchSre | 56.2 | 1.1 |
| OmniScience Accuracy | 59.4 | 24.3 |
| OmniScience Non-Hallucination | 10.6 | 13 |
| scicode | 56.9 | 42 |
| TauBench V3 - Banking | 44.3 | 10.3 |
| Terminal-Bench 2.1 | 89.5 | 38.6 |
| Terminal-Bench Hard | 65.9 | 28.8 |
| τ²-Bench Telecom (AA run) | 85.1 | 67.8 |
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.