Goal-directed behavior and habit are parallel and, at times, competing processes. The relative balance of flexible, goal-directed responding as compared to inflexible habitual responding is highly dependent on experience (e.g., training history in a task) and conditions under which the behavior was formed. Reinforcer devaluation tasks have been used widely across species to study the neurobiology of goal-directed behavior. In rodents, under some conditions, extended training in reinforcer devaluation tasks transforms goal-directed responses into habits, rendering the animals insensitive to devaluation. In nonhuman primates, no studies have previously evaluated the impact of extended training. Here we trained four macaques in a variant of the standard reinforcer devaluation task (Málková et al., 1997), in which we presented objects with either a standard number of exposures (i.e., up to 55) or with a high number of exposures (i.e., up to 454). We tested for goal-directed behavior at three time points during the course of this extended training with different combinations of high- and low-repetition objects and stratified results based on whether the preferred or nonpreferred reinforcer was devalued. We found robust devaluation effects across all three cycles of training; however, the magnitude of the effect was modulated by reinforcer preference and by the relative training history of the objects. These data argue against habit formation after overtraining in the reinforcer devaluation task in macaques, a finding that is consistent with reports in humans and with tasks in rodents that employ multiple stimuli, reinforcers, and instrumental actions. (PsycInfo Database Record (c) 2022 APA, all rights reserved).