Across 15 shared benchmarks, Gemini 1.5 Pro scores higher on 2 and Qwen3.5 122B A10B on 13. The widest gap is HLE, where Qwen3.5 122B A10B scores 47.5 against 4.9.
| Benchmark | Gemini 1.5 Pro | Qwen3.5 122B A10B |
|---|---|---|
| AA Intelligence | 10 | 32.8 |
| AI2D | 80.9 | 93.3 |
| arena_vision | 1179 | 1246 |
| Artificial Analysis Coding Index | 23.6 | 45.7 |
| GPQA Diamond | 59.1 | 86.6 |
| GSM8K | 95.2 | 94.5 |
| HLE | 4.9 | 47.5 |
| ifeval | 89.4 | 93.4 |
| MathVista | 70.6 | 87.4 |
| MMLU-Pro | 75.8 | 86.7 |
| MMMU | 68.4 | 83.9 |
| MMMU-Pro | 55 | 76.9 |
| OCRBench | 800 | 92.1 |
| scicode | 29.5 | 42 |
| SWE-bench Verified | 34.2 | 72 |
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.