Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Gemini 3 Pro | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 91.9 | 90.5 |
| Humanity's Last Exam | 45.8 | 36.4 |
| ARC-AGI v2 | 31.1 | — |
math
| Benchmark | Gemini 3 Pro | Kimi K2.6 |
|---|---|---|
| AIME 2025 | 100 | — |
coding
| Benchmark | Gemini 3 Pro | Kimi K2.6 |
|---|---|---|
| SWE-bench Verified | 76.2 | 80.2 |
| SWE-Bench Pro | — | 58.6 |
| SciCode | — | 52.2 |
multimodal
| Benchmark | Gemini 3 Pro | Kimi K2.6 |
|---|---|---|
| MMMU-Pro | 81 | 80.1 |
| CharXiv Reasoning | 81.4 | 86.7 |
| ScreenSpot Pro | 72.7 | — |
agent
| Benchmark | Gemini 3 Pro | Kimi K2.6 |
|---|---|---|
| BrowseComp | — | 86.3 |
| Toolathlon | — | 50 |
| APEX Agents | — | 27.9 |
Category Winners
coding
74.3
math
84.0
reasoning
M
Kimi K2.675.1
general
65.2
multimodal
M
Kimi K2.681.7
agent
M
Kimi K2.653.8