| Benchmark | Gemini 3.6 Flash | GPT-4o mini |
|---|---|---|
| AA Agentic Index | 40.5 | 1 |
| AA Intelligence | 52 | 7 |
| ARC-AGI-2 | 2.6 | 0 |
| Artificial Analysis Coding Index | 69.2 | 11.4 |
| gdpval | 46.1 | 0 |
| GPQA Diamond | 92.8 | 42.6 |
| HLE | 40.8 | 4.2 |
| MMMU-Pro | 83.2 | 41.5 |
| scicode | 52.7 | 22.9 |
| TauBench V3 - Banking | 29.9 | 2.9 |
| Terminal-Bench 2.1 | 78 | 5.6 |
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.