| Benchmark | Gemini 3 Flash Preview | o1 |
|---|---|---|
| AA Agentic Index | 49.7 | 31.1 |
| AA Intelligence | 38.7 | 24 |
| AA-LCR | 73 | 63.3 |
| AA-Omniscience | 10.1 | -10.6 |
| AIME 2025 | 99.7 | 81.7 |
| arena_vision | 1260 | 1193 |
| Artificial Analysis Coding Index | 42.6 | 39.7 |
| critpt | 8.6 | 0.3 |
| gdpval | 35.2 | 11.5 |
| GPQA Diamond | 90.4 | 75.7 |
| harmbench | 70 | 96.3 |
| HLE | 43.5 | 7.7 |
| IFBench | 78 | 70.3 |
| livebench | 72.4 | 52.3 |
| mmmlu | 91.8 | 87.7 |
| OmniScience Accuracy | 53.4 | 34.7 |
| OmniScience Non-Hallucination | 7.6 | 30.7 |
| scicode | 50.6 | 35.8 |
| simplebench | 61.1 | 41.7 |
| simpleqa | 68.7 | 47 |
| SWE-bench Verified | 78 | 48.9 |
| Terminal-Bench Hard | 38.6 | 12.9 |
| τ²-Bench Telecom (AA run) | 80.4 | 62.6 |
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.