In 2007, a study carried out by Hamlin, Wynn, and Bloom provided concrete evidence that infants as young as 6 months were capable of social evaluation, displaying an early preference for agents performing a prosocial behavior. Since then the development of early social abilities to judge other's behavior has been the topic of a growing body of research. The present paper reviews studies conducted between 2007 and 2015 that experimentally examined infants' social evaluation abilities by testing their preference for agents acting prosocially. We performed a detailed analysis of a corpus of 16 research studies including 59 experimental results, scrutinizing their methods and findings, and identifying their convergent and divergent features. This analysis showed that a preference for agents who perform prosocial behaviors (as opposed to antisocial or neutral) was present in a majority of infants, but some conflicting results have also been reported. The rich interpretation that infants are endowed with mature socio-moral evaluation abilities has not really been sufficiently discussed. In order to deepen this debate, we assessed other studies that have further explored infants' understanding of the social value of behaviors. Many of the studies provide evidence that young infants manage to identify and prefer the prosocial agent by taking into account the context and agents' mental states beyond the behavior itself. In this study two specific areas are assessed: (1) studies that have previously explored social evaluation abilities beyond a basic preference for prosocial behavior and (2) current theories which attempt to explain how and why such preferences could exist so early in infancy. Future directions for research on social evaluation abilities in infants are also discussed as well as a review of the literature.