| Benchmark | Gemini 3 Pro | Nemotron 3 Super 120B A12B |
|---|---|---|
| AA Agentic Index | 52 | 8.8 |
| AA Intelligence | 40.6 | 25.7 |
| AA-LCR | 73 | 60.3 |
| AA-Omniscience | 15.3 | -41.5 |
| AIME 2025 | 100 | 90.2 |
| aime_2026 | 91.7 | 91.7 |
| AIME25 no tools | 96 | 92.2 |
| apexAgents | 18.4 | 1.8 |
| Artificial Analysis Coding Index | 46.5 | 37.7 |
| browsecomp | 59.2 | 31.3 |
| critpt | 9.1 | 3.1 |
| gdpval | 34.2 | 9.9 |
| GPQA Diamond | 91.9 | 82.7 |
| HLE | 45.8 | 22.8 |
| hmmt_feb_2026 | 86.4 | 85.2 |
| IFBench | 70.4 | 72.6 |
| LiveCodeBench | 92 | 82.1 |
| MMLU-Pro | 90.1 | 83.8 |
| multichallenge | 65.7 | 55.2 |
| OmniScience Accuracy | 55.8 | 24.3 |
| OmniScience Non-Hallucination | 10 | 13 |
| scicode | 56.1 | 42 |
| SWE-bench Multilingual | 68.7 | 45.8 |
| SWE-bench Verified | 78 | 53.7 |
| Terminal-Bench 2.0 | 54.2 | 31 |
| Terminal-Bench Hard | 56.9 | 28.8 |
| τ²-Bench Telecom (AA run) | 98 | 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.