Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Gemini 3.1 Pro | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 94.3 | 90.5 |
| Humanity's Last Exam | 51.4 | 36.4 |
| ARC-AGI v2 | 77.1 | — |
coding
| Benchmark | Gemini 3.1 Pro | Kimi K2.6 |
|---|---|---|
| SWE-bench Verified | 80.6 | 80.2 |
| SWE-Bench Pro | 54.2 | 58.6 |
| SciCode | 59 | 52.2 |
multimodal
| Benchmark | Gemini 3.1 Pro | Kimi K2.6 |
|---|---|---|
| MMMU-Pro | 80.5 | 80.1 |
| CharXiv Reasoning | — | 86.7 |
agent
| Benchmark | Gemini 3.1 Pro | Kimi K2.6 |
|---|---|---|
| BrowseComp | 85.9 | 86.3 |
| MCP Atlas | 69.2 | — |
| APEX Agents | 33.5 | 27.9 |
| Toolathlon | — | 50 |
Category Winners
coding
M
Kimi K2.671.2
reasoning
89.6
general
53.5
multimodal
82.2
agent
69.1