| Benchmark | Gemini 3.5 Flash | Llama 3.1 Instruct 405B |
|---|---|---|
| AA Agentic Index | 70.4 | 6.3 |
| AA Intelligence | 52 | 8.5 |
| AA-LCR | 81 | 24.7 |
| AA-Omniscience | 21.2 | -17.1 |
| Artificial Analysis Coding Index | 70.1 | 14.5 |
| critpt | 13.1 | 0 |
| gdpval | 57.8 | 0 |
| GPQA Diamond | 92.2 | 51.5 |
| HLE | 42.7 | 4.2 |
| IFBench | 76.3 | 39 |
| OmniScience Accuracy | 51.4 | 23.2 |
| OmniScience Non-Hallucination | 38.2 | 49 |
| scicode | 53.1 | 29.9 |
| simplebench | 76.7 | 23 |
| SWE-bench Pro | 55.1 | 11.2 |
| Terminal-Bench Hard | 46.2 | 6.8 |
| τ²-Bench Telecom (AA run) | 95.6 | 19 |
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.