Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-4.1 nano | Qwen3 VL 8B Instruct |
|---|---|---|
| GPQA Diamond | 50.3 | — |
multimodal
| Benchmark | GPT-4.1 nano | Qwen3 VL 8B Instruct |
|---|---|---|
| MMMU | 55.4 | — |
| CharXiv Reasoning | 40.5 | 46.4 |
| MMMU-Pro | — | 55.9 |
| ScreenSpot Pro | — | 54.6 |
general
| Benchmark | GPT-4.1 nano | Qwen3 VL 8B Instruct |
|---|---|---|
| MMMLU | 66.9 | — |
agent
| Benchmark | GPT-4.1 nano | Qwen3 VL 8B Instruct |
|---|---|---|
| TAU-Bench Retail | 22.6 | — |
| OSWorld | — | 33.9 |
math
| Benchmark | GPT-4.1 nano | Qwen3 VL 8B Instruct |
|---|---|---|
| AIME 2025 | — | 45.9 |
Category Winners
math
Q
Qwen3 VL 8B Instruct5.5
reasoning
19.2
general
4.6
multimodal
Q
Qwen3 VL 8B Instruct15.8
agent
Q
Qwen3 VL 8B Instruct28.8