Social separation is an increasingly used strategy to experimentally assess psychogenic stress in nonhuman primates, yet it is also inherently required for captive management. Surprisingly, little is known on the effects of an individual left alone in a highly familiar place, compared to novel-environment isolation. Therefore, using stable heterosexual pairs of adult marmosets (Callithrix penicillata), we assessed in the individual left behind in the home-cage: (1) the behavior during and after a 5 and 15 min separation; (2) the behavior and cortisol levels before, during and after a 7-day separation; and (3) the relationship between these two components of the stress response. All conditions induced social contact calls and hyperlocomotion, with prolonged separation yielding a habituation effect (p < 0.05). How each parameter returned to baseline following the reunion depended on the duration of the preceding separation interval, with post-reunion affiliative behaviors increasing significantly only after longer intervals (experiment 1: p = 0.02; experiment 2: p = 0.001). Cortisol concentrations, which also increased during the 7-day separation (p = 0.04), were temporally dissociated from and not correlated with the marmosets’ behavioral response (p > 0.05). Interestingly, cortisol levels during the social separation phase were negatively correlated with post-reunion affiliative behavior (p = 0.05). Stress-induced cortisol may thus hinder putative behavioral coping strategies. However, post-reunion affiliative behavior was not correlated with the last cortisol measure (p = 0.21). Therefore, social disruption in a familiar environment – even if brief – induces stress-related responses and behavioral coping strategies. These findings bear important implication for captive management and research protocols using marmosets.