| Benchmark | Gemini 3.5 Flash | Kimi K2 (Reasoning) |
|---|---|---|
| AA Agentic Index | 70.4 | 47.9 |
| AA Intelligence | 52 | 33.5 |
| AA-LCR | 81 | 70.3 |
| AA-Omniscience | 21.2 | -21.4 |
| Artificial Analysis Coding Index | 70.1 | 34.8 |
| critpt | 13.1 | 2.6 |
| frontiermath_tier_4 | 14.6 | 0 |
| gdpval | 57.8 | 24.5 |
| GPQA Diamond | 92.2 | 84.5 |
| HLE | 42.7 | 23.9 |
| IFBench | 76.3 | 68.1 |
| OmniScience Accuracy | 51.4 | 30.9 |
| OmniScience Non-Hallucination | 38.2 | 25.8 |
| scicode | 53.1 | 44.8 |
| simplebench | 76.7 | 39.6 |
| Terminal-Bench 2.0 | 76.2 | 35.7 |
| Terminal-Bench Hard | 46.2 | 31.1 |
| τ²-Bench Telecom (AA run) | 95.6 | 93 |
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.