Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5 nano | LongCat-Flash-Lite |
|---|---|---|
| GPQA Diamond | 71.2 | 66.8 |
| Humanity's Last Exam | 8.7 | — |
math
| Benchmark | GPT-5 nano | LongCat-Flash-Lite |
|---|---|---|
| AIME 2025 | 85.2 | 63.2 |
| FrontierMath | 9.6 | — |
coding
| Benchmark | GPT-5 nano | LongCat-Flash-Lite |
|---|---|---|
| SWE-bench Verified | — | 54.4 |
| TerminalBench | — | 33.8 |
Category Winners
coding
M
LongCat-Flash-Lite36.0
math
39.0
reasoning
M
LongCat-Flash-Lite50.2