Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
agent
| Benchmark | Claude Code | GPT-5.5 Pro |
|---|---|---|
| GAIA | 72 | — |
| WebArena | 35 | — |
| BrowseComp | — | 90.1 |
coding
| Benchmark | Claude Code | GPT-5.5 Pro |
|---|---|---|
| SWE-bench Verified | 72.7 | — |
reasoning
| Benchmark | Claude Code | GPT-5.5 Pro |
|---|---|---|
| Humanity's Last Exam | — | 57.2 |
math
| Benchmark | Claude Code | GPT-5.5 Pro |
|---|---|---|
| FrontierMath | — | 39.6 |
Category Winners
coding
67.2
math
57.5
reasoning
93.4
agent
89.6