Neura Intelligence Index
AI Model Benchmarks
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 91.9 | 87.6 |
| Humanity's Last Exam | 45.8 | 50.2 |
| ARC-AGI v2 | 31.1 | — |
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| AIME 2025 | 100 | 96.1 |
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 76.2 | 76.8 |
| SWE-Bench Pro | — | 50.7 |
| SciCode | — | 48.7 |
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| MMMU-Pro | 81 | 78.5 |
| CharXiv Reasoning | 81.4 | 77.5 |
| ScreenSpot Pro | 72.7 | — |
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| BrowseComp | — | 74.9 |