This study aimed to investigate the effects of gender pairings on collaborative problem-solving performance, processes, and attitudes in a social learning context. Three types of pairings (i.e., male-male, female-female, and mixed pairings) were considered in an empirical study with 222 tenth-grade students. The selection of three different schools facilitated discussions regarding which schools were more divergent and competitive in a social learning context. The students were asked to solve computer science problems on a social media platform. The results revealed that (1) the single-gender groups had more focused discussions than the mixed-gender groups. Specifically, the male-male groups tended to develop and test their solutions directly without spending significant time on problem identification. Consequently, the single-gender groups exhibited superior performance compared to the mixed-gender groups in terms of applying their knowledge to problem solving. In terms of attitudes toward social learning, the female-female groups were more attentive to the benefits of social learning than the male-male groups. (2) The mixed gender groups had more diverse and divergent discussions compared to the single-gender groups. The educational implications of these findings are also discussed in this paper.