The study aimed to examine the difference between cheating and persistence during a persistence task to advance persistence measurement. Through a within-subject design (N = 78, mean age: 5.2 years), two different versions of the puzzle box task were administered. The original puzzle box task was administered in condition I (i.e., open version). Children could exhibit three types of behavior: persistence, cheating, and off-task. In condition II (i.e., the fixed version), the box was manipulated to make it impossible to cheat. Hence, children could only exhibit persistence or off-task behavior. The results revealed that while persistence remained stable across the two conditions, the amount of off-task increased substantially when there was no possibility to cheat. These findings indicate that persistence and cheating can be reliably distinguished and provide an empirical basis for analyzing behavioral dynamics between persistence, cheating, and off-task in the puzzle box task.