Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| SWE-Bench Pro | 56.2 | — |
agent
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| Toolathlon | 46.3 | — |
| OSWorld | — | 30.3 |
reasoning
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| GPQA Diamond | — | 70.4 |
math
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| AIME 2025 | — | 69.3 |
multimodal
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| MMMU-Pro | — | 60.4 |
| CharXiv Reasoning | — | 48.9 |
| ScreenSpot Pro | — | 60.5 |
general
| Benchmark | MiniMax M2.7 | Qwen3 VL 30B A3B Instruct |
|---|---|---|
| SimpleQA | — | 27 |
Category Winners
coding
38.6
math
Q
Qwen3 VL 30B A3B Instruct32.0
reasoning
Q
Qwen3 VL 30B A3B Instruct57.4
general
Q
Qwen3 VL 30B A3B Instruct33.3
multimodal
Q
Qwen3 VL 30B A3B Instruct26.4
agent
41.6