Neura Intelligence Index
AI Model Benchmarks
| Benchmark | Claude Opus 4.6 | Claude Sonnet 4.5 |
|---|---|---|
| GPQA Diamond | 91.3 | 83.4 |
| Humanity's Last Exam | 53.1 | — |
| ARC-AGI v2 | 68.8 | — |
| Benchmark | Claude Opus 4.6 | Claude Sonnet 4.5 |
|---|---|---|
| AIME 2025 | 99.8 | 87 |
| Benchmark | Claude Opus 4.6 | Claude Sonnet 4.5 |
|---|---|---|
| SWE-bench Verified | 80.8 | — |
| TerminalBench | — | 50 |
| Benchmark | Claude Opus 4.6 | Claude Sonnet 4.5 |
|---|---|---|
| MMMU-Pro | 77.3 | — |
| CharXiv Reasoning | 77.4 | — |
| Benchmark | Claude Opus 4.6 | Claude Sonnet 4.5 |
|---|---|---|
| OSWorld | 72.7 | 61.4 |
| BrowseComp | 84 | — |
| MCP Atlas | 62.7 | — |
| TAU-Bench Retail | — | 86.2 |