| Benchmark | DeepSeek-V3.2 | Gemini 1.5 Pro |
|---|---|---|
| AA Intelligence | 32.8 | 10 |
| ARC-AGI-2 | 4 | 0.8 |
| Artificial Analysis Coding Index | 44.2 | 23.6 |
| GPQA Diamond | 84 | 59.1 |
| HLE | 40.8 | 4.9 |
| LiveCodeBench | 86 | 30.5 |
| MMLU-Pro | 86 | 75.8 |
| mrcr | 55.5 | 82.6 |
| scicode | 39 | 29.5 |
| SWE-bench Verified | 73.1 | 34.2 |
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.