We examined whether the moderating role of cortisol in the relation between physical peer victimization and physical aggression was better accounted for by a diathesis-stress model or a differential susceptibility model using a multi-informant approach (direct observations, teacher reports, and parent reports) of 198 preschool-aged children attending high-quality child care. Controlling for the influence of household income, hours per week in child care, and child age, we found that our results supported a differential susceptibility effect for boys but not for girls. This effect was replicated within reporters (observer and parent reports) and across reporters (parent-reported victimization and teacher-rated aggression) but for boys only. At higher levels of peer victimization, lower levels of basal cortisol were associated with higher levels of physical aggression, but, at lower levels of peer victimization, lower levels of basal cortisol were associated with lower levels of physical aggression. Furthermore, at higher levels of peer victimization, higher levels of basal cortisol were associated with lower levels of physical aggression, but, at lower levels of peer victimization, higher levels of basal cortisol were associated with higher levels of physical aggression. These results highlight the complex interplay between the social environment and biobehavioral systems in early childhood and the value of considering a differential susceptibility framework in peer-relations research.