Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Haiku 4.5 | LongCat-Flash-Thinking |
|---|---|---|
| GPQA Diamond | 73 | 81.5 |
math
| Benchmark | Claude Haiku 4.5 | LongCat-Flash-Thinking |
|---|---|---|
| AIME 2025 | 80.7 | 90.6 |
coding
| Benchmark | Claude Haiku 4.5 | LongCat-Flash-Thinking |
|---|---|---|
| SWE-bench Verified | 73.3 | 59.4 |
| TerminalBench | 41 | — |
general
| Benchmark | Claude Haiku 4.5 | LongCat-Flash-Thinking |
|---|---|---|
| MMMLU | 83 | — |
agent
| Benchmark | Claude Haiku 4.5 | LongCat-Flash-Thinking |
|---|---|---|
| OSWorld | 50.7 | — |
Category Winners
coding
70.3
math
M
LongCat-Flash-Thinking70.9
reasoning
M
LongCat-Flash-Thinking75.4
general
48.8
agent
57.3