Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Kimi K2-Thinking-0905 | o1 |
|---|---|---|
| GPQA Diamond | 84.5 | 78 |
| Humanity's Last Exam | 51 | — |
| BigBench-Hard | — | 91.5 |
coding
| Benchmark | Kimi K2-Thinking-0905 | o1 |
|---|---|---|
| SWE-bench Verified | 71.3 | 48.9 |
| TerminalBench | 47.1 | — |
| SciCode | 44.8 | — |
| HumanEval | — | 92.4 |
agent
| Benchmark | Kimi K2-Thinking-0905 | o1 |
|---|---|---|
| BrowseComp | 60.2 | — |
general
| Benchmark | Kimi K2-Thinking-0905 | o1 |
|---|---|---|
| MMLU-Pro | — | 79.2 |
| IFEval | — | 86.3 |
| Chatbot Arena Elo | — | 1350 |
Category Winners
coding
M
Kimi K2-Thinking-090566.6
math
94.2
reasoning
M
Kimi K2-Thinking-090584.5
general
86.7
agent
M
Kimi K2-Thinking-090541.4