Neura Intelligence Index
AI Model Benchmarks
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| GPQA Diamond | 89.9 | 84.5 |
| Humanity's Last Exam | 49 | 51 |
| ARC-AGI v2 | 58.3 | — |
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| SWE-bench Verified | 79.6 | 71.3 |
| TerminalBench | — | 47.1 |
| SciCode | — | 44.8 |
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| MMMU-Pro | 75.6 | — |
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| MMMLU | 89.3 | — |
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| OSWorld | 72.5 | — |
| BrowseComp | 74.7 | 60.2 |
| MCP Atlas | 61.3 | — |
| Benchmark | Claude Sonnet 4.6 | Kimi K2-Thinking-0905 |
|---|---|---|
| AIME 2025 | — | 100 |