Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | DeepSeek R1 Zero | LongCat-Flash-Thinking-2601 |
|---|---|---|
| GPQA Diamond | 73.3 | 80.5 |
| Humanity's Last Exam | — | 25.2 |
math
| Benchmark | DeepSeek R1 Zero | LongCat-Flash-Thinking-2601 |
|---|---|---|
| AIME 2025 | — | 99.6 |
coding
| Benchmark | DeepSeek R1 Zero | LongCat-Flash-Thinking-2601 |
|---|---|---|
| SWE-bench Verified | — | 70 |
agent
| Benchmark | DeepSeek R1 Zero | LongCat-Flash-Thinking-2601 |
|---|---|---|
| BrowseComp | — | 56.6 |
Category Winners
coding
M
LongCat-Flash-Thinking-260160.5
math
M
LongCat-Flash-Thinking-260183.8
reasoning
60.8
agent
M
LongCat-Flash-Thinking-260133.4