Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Fable 5 | DeepSeek-V4-Pro-0813 |
|---|---|---|
| Humanity's Last Exam | 64.5 | 60 |
coding
| Benchmark | Claude Fable 5 | DeepSeek-V4-Pro-0813 |
|---|---|---|
| SWE-bench Verified | 95 | — |
| SWE-Bench Pro | 80 | — |
agent
| Benchmark | Claude Fable 5 | DeepSeek-V4-Pro-0813 |
|---|---|---|
| Toolathlon | — | 74.1 |
Category Winners
coding
98.1
reasoning
96.6
agent
96.9