Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-4 Turbo | Qwen2.5 VL 32B Instruct |
|---|---|---|
| GPQA Diamond | 48 | 46 |
multimodal
| Benchmark | GPT-4 Turbo | Qwen2.5 VL 32B Instruct |
|---|---|---|
| MMMU | — | 70 |
| MMMU-Pro | — | 49.5 |
| ScreenSpot Pro | — | 39.4 |
agent
| Benchmark | GPT-4 Turbo | Qwen2.5 VL 32B Instruct |
|---|---|---|
| OSWorld | — | 5.9 |
Category Winners
reasoning
17.4
multimodal
Q
Qwen2.5 VL 32B Instruct23.8
agent
Q
Qwen2.5 VL 32B Instruct3.2