Group cohesion is thought to be beneficial for the mental health of group members. However, this association could depend on whether the focus lies on one’s perceived personal relationship to other group members or on one’s perception of cohesion of the whole group. The association between higher perceived group cohesion of the whole group and improved mental health might decrease if individual group members are exposed to interpersonal mistreatment, such as harassment. Using γ regressions, we examined the interaction between perceived cohesion of the whole group and perceived harassment predicting depressive and anxiety symptoms. Twelve months following military deployment, 1,483 participants rated harassment and cohesion within their units during the time of deployment and reported on anxiety and depressive symptoms in the past 7 days. Cohesion was negatively related to anxiety, χ2(2) = 27.12, p <.001, and depressive symptoms, χ2(2) = 31.07, p <.001. However, with increasing reported harassment, these associations decreased, anxiety symptoms: χ2(2) = 7.47, p =.024, depressive symptoms: χ2(2) = 10.73, p =.005. The negative association between perceived group cohesion and internalizing symptoms is in line with existing findings. However, this is the first study to provide preliminary evidence for the notion that individuals who are exposed to harassment benefit less from perceived group cohesion.