Agentic Execution on Kimi CLI 74 paired tasks
22Test Statistic (t/W)Wilcoxon signed-rank
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Wilcoxon signed-rankTest Type=Wilcoxon signed-rank2026.05 | 22 | 0.005 | 0.01 | |
| Non-tie only (n = 17)Test Type=Non-tie only, n=172026.05 | 3.449 | 0.0033 | 0.01 | |
| Paired t-testTest Type=Paired t-test2026.05 | 2.815 | 0.0063 | 0.01 | |
| Cohen’s d (paired)Test Type=Effect Size2026.05 | 0.327 | — | — |