Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | DeepSeek-V4-Flash-Max | Qwen3.6-27B |
|---|---|---|
| GPQA Diamond | 88.1 | 87.8 |
| Humanity's Last Exam | 45.1 | 24 |
coding
| Benchmark | DeepSeek-V4-Flash-Max | Qwen3.6-27B |
|---|---|---|
| SWE-bench Verified | 79 | 77.2 |
| SWE-Bench Pro | 52.6 | 53.5 |
general
| Benchmark | DeepSeek-V4-Flash-Max | Qwen3.6-27B |
|---|---|---|
| SimpleQA | 34.1 | — |
agent
| Benchmark | DeepSeek-V4-Flash-Max | Qwen3.6-27B |
|---|---|---|
| BrowseComp | 73.2 | — |
| Toolathlon | 47.8 | — |
| MCP Atlas | 69 | — |
multimodal
| Benchmark | DeepSeek-V4-Flash-Max | Qwen3.6-27B |
|---|---|---|
| MMMU | — | 82.9 |
| MMMU-Pro | — | 75.8 |
| CharXiv Reasoning | — | 78.4 |
Category Winners
coding
Q
Qwen3.6-27B55.1
reasoning
80.5
general
44.0
multimodal
Q
Qwen3.6-27B72.2
agent
48.4