Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Kimi K2-Instruct-0905 | MiniMax M1 40K |
|---|---|---|
| GPQA Diamond | 75.1 | 69.2 |
| Humanity's Last Exam | 4.7 | 7.2 |
math
| Benchmark | Kimi K2-Instruct-0905 | MiniMax M1 40K |
|---|---|---|
| AIME 2025 | 49.5 | 74.6 |
coding
| Benchmark | Kimi K2-Instruct-0905 | MiniMax M1 40K |
|---|---|---|
| SWE-bench Verified | 65.8 | 55.6 |
| TerminalBench | 25 | — |
general
| Benchmark | Kimi K2-Instruct-0905 | MiniMax M1 40K |
|---|---|---|
| SimpleQA | 31 | 17.9 |
agent
| Benchmark | Kimi K2-Instruct-0905 | MiniMax M1 40K |
|---|---|---|
| TAU-Bench Retail | — | 67.8 |
Category Winners
coding
M
Kimi K2-Instruct-090534.4
math
41.5
reasoning
M
Kimi K2-Instruct-090536.9
general
M
Kimi K2-Instruct-090539.1
agent
50.1