Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Gemini 1.5 Pro | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| GPQA Diamond | 59.1 | 70.4 |
multimodal
| Benchmark | Gemini 1.5 Pro | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| MMMU | 65.9 | — |
| MMMU-Pro | — | 60.4 |
| CharXiv Reasoning | — | 48.9 |
| ScreenSpot Pro | — | 60.5 |
math
| Benchmark | Gemini 1.5 Pro | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| AIME 2025 | — | 69.3 |
general
| Benchmark | Gemini 1.5 Pro | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| SimpleQA | — | 27 |
agent
| Benchmark | Gemini 1.5 Pro | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| OSWorld | — | 30.3 |
Category Winners
math
Q
Qwen3 VL 30B A3B Instruct31.4
reasoning
Q
Qwen3 VL 30B A3B Instruct54.6
general
Q
Qwen3 VL 30B A3B Instruct33.4
multimodal
43.9
agent
Q
Qwen3 VL 30B A3B Instruct22.1