Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Opus 5 | DeepSeek-V4-Pro-0813 |
|---|---|---|
| Humanity's Last Exam | 64.7 | 60 |
agent
| Benchmark | Claude Opus 5 | DeepSeek-V4-Pro-0813 |
|---|---|---|
| BrowseComp | 90.8 | — |
| Toolathlon | — | 74.1 |
Category Winners
reasoning
96.5
agent
93.3