Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| GPQA Diamond | 84.2 | 85.5 |
| Humanity's Last Exam | — | 48.5 |
math
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| AIME 2025 | 97 | — |
coding
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| SWE-bench Verified | 73.5 | 72.4 |
| SWE-Bench Pro | 52.8 | — |
multimodal
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| MMMU | — | 82.3 |
| MMMU-Pro | — | 75 |
| CharXiv Reasoning | — | 79.5 |
| ScreenSpot Pro | — | 70.3 |
general
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| MMMLU | — | 85.9 |
agent
| Benchmark | MAI-Thinking-1 | Qwen3.5-27B |
|---|---|---|
| BrowseComp | — | 61 |
Category Winners
coding
Q
Qwen3.5-27B66.1
math
80.6
reasoning
Q
Qwen3.5-27B81.6
general
Q
Qwen3.5-27B60.8
multimodal
Q
Qwen3.5-27B70.0
agent
Q
Qwen3.5-27B41.7