Social interaction is a core aspect of mammalian behavior, and alterations in social behaviors are found across many neurodevelopmental and psychiatric conditions. Mice display a range of socioemotional behaviors and are commonly used as models to investigate the neuronal circuits and molecular mechanisms underlying differences in social behaviors. Analyzing social interactions in mice is often done by manual quantification of videos. Although multiple methods for automated tracking have been developed, reliable tracking and automated behavioral classification of multiple freely-moving unmarked animals have remained challenging. The use of an unbiased classification system in a more naturalistic environment could help obtain a translatable way to study social behaviors, in particular in mouse models for neuropsychiatric conditions. Here, we use open-source toolkits: DeepLabCut, and DeepOF to track and quantify reciprocal self-selected social interactions in pairs of freely moving sex-and age-matched animals. In addition, we use the LiveMouseTracker (LMT) to investigate home cage phenotyping and examines how this compares with classical behavioral phenotyping in female and male mice that lack the autism-associated gene Shank3. We show that DeepLabCut can track the movement of two freely interacting size-matched juvenile mice, and behavioral classification using supervised methods can detect differences in the social interaction profiles between Shank3 knockout and control mice. Quantifying behavior in an unbiased way remains a challenge in animal research. We confirm that freely-available open-source toolkits can be used to track and classify social interactions in a home cage, thereby providing a simple, low-cost solution to analyze social behaviors in age-and sex-matched mice.