Virtual peer interaction is prevalent among adolescents (Anderson & Jiang, 2018), but little is known about how adolescents’ virtual interactions with peers compare to their in-person interactions. The present study aimed to compare adolescents’ in-person and virtual interactions in a multiplayer video game during an initial interaction with an unfamiliar peer to examine differences in social behavior, physiological responding, and perceptions of interaction quality. The study also aimed to investigate how motivations for solitude related to interaction quality, and whether these associations differed across virtual and in-person interaction. Participants were 72 adolescents (78% male, Mage = 12.49) from the Washington, DC metropolitan area who interacted with an unfamiliar peer in the lab using the multiplayer game Minecraft. Pairs of participants were randomly assigned to interact with one another in-person, sitting in the same room next to each other, or virtually, able to communicate using the text-based chat feature. Participants completed questionnaires about their motivations for solitude prior to the interaction. They also completed questionnaires about their self-perceptions and affect before and after the interaction, as well as their perceptions of the interaction quality after the interaction. Participants’ social engagement and their social initiations and the partner’s responses were observed during the interaction, and their respiratory sinus arrhythmia was measured before, during, and after the interaction. Results showed that quantity of social interaction was higher in the in-person condition, but perceived quality of the interaction was higher in the virtual condition. Participants spent more time communicating with one another and made more social initiations in the in-person condition. However, participants in the virtual condition received more successful responses to their social initiations and reported enjoying the interaction marginally more, feeling less passive and more assertive, and viewing themselves as more socially competent following the interaction. Participants’ physiological responding did not differ across conditions. Shyness was related to less positive emotional responses to the interaction, particularly in the in-person condition, while other motivations for solitude were less consistently related to social difficulties. These findings suggest that virtual interaction can be an engaging context that facilitates high-quality interactions between unfamiliar peers, and it may be particularly helpful for shy adolescents.