Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | o3 | Qwen3.6-35B-A3B |
|---|---|---|
| GPQA Diamond | 83.3 | 86 |
| Humanity's Last Exam | 14.7 | 21.4 |
| ARC-AGI v2 | 6.5 | — |
math
| Benchmark | o3 | Qwen3.6-35B-A3B |
|---|---|---|
| AIME 2025 | 86.4 | — |
| FrontierMath | 15.8 | — |
coding
| Benchmark | o3 | Qwen3.6-35B-A3B |
|---|---|---|
| SWE-bench Verified | 69.1 | 73.4 |
| SWE-Bench Pro | — | 49.5 |
multimodal
| Benchmark | o3 | Qwen3.6-35B-A3B |
|---|---|---|
| MMMU | 82.9 | 81.7 |
| MMMU-Pro | 76.4 | 75.3 |
| CharXiv Reasoning | 78.6 | 78 |
agent
| Benchmark | o3 | Qwen3.6-35B-A3B |
|---|---|---|
| BrowseComp | 49.7 | — |
| Toolathlon | — | 26.9 |
| MCP Atlas | — | 62.8 |
Category Winners
coding
59.1
math
43.4
reasoning
Q
Qwen3.6-35B-A3B56.9
multimodal
75.1
agent
23.1