Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| GPQA Diamond | 81.4 | 88.4 |
| Humanity's Last Exam | — | 28.7 |
math
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| AIME 2025 | 92.3 | — |
coding
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| SWE-bench Verified | 70 | 76.4 |
| TerminalBench | 41.3 | — |
multimodal
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| MMMU-Pro | 63.5 | — |
general
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| MMMLU | — | 88.5 |
agent
| Benchmark | Nova 2 Pro | Qwen3.5-397B-A17B |
|---|---|---|
| BrowseComp | — | 69 |
| Toolathlon | — | 38.3 |
Category Winners
coding
Q
Qwen3.5-397B-A17B74.6
math
Nova 2 Pro73.7
reasoning
Nova 2 Pro75.2
general
Q
Qwen3.5-397B-A17B70.8
multimodal
Nova 2 Pro37.0
agent
Q
Qwen3.5-397B-A17B36.7