Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 80.9 | 87.6 |
| Humanity's Last Exam | — | 50.2 |
math
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| AIME 2025 | 78 | 96.1 |
coding
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 74.5 | 76.8 |
| TerminalBench | 43.3 | — |
| SWE-Bench Pro | — | 50.7 |
| SciCode | — | 48.7 |
general
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| MMMLU | 89.5 | — |
agent
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| TAU-Bench Retail | 82.4 | — |
| BrowseComp | — | 74.9 |
multimodal
| Benchmark | Claude Opus 4.1 | Kimi K2.5 |
|---|---|---|
| MMMU-Pro | — | 78.5 |
| CharXiv Reasoning | — | 77.5 |
Category Winners
coding
75.2
math
M
Kimi K2.579.1
reasoning
M
Kimi K2.586.0
general
73.8
multimodal
M
Kimi K2.570.6
agent
86.7