Historically, it was widely assumed that males dominate females socially in most mammals. However, recent studies revealed significant variation within and among species, opening new possibilities to explore the extent and drivers of sex biases in dominance relations. This study uses quantitative data from 253 populations across 121 primate species to investigate the distribution of, and factors associated with, sex biases in the outcome of male–female contests. We first showed that male–female contests are common (around half of all contests) and that males win >90% of these contests in less than 20% of populations. We next tested five hypotheses to explain sex biases in dominance relations. We found that female-biased dominance primarily occurs in primate societies where females have substantial reproductive control, as in monogamous, sexually monomorphic, and arboreal species. Female-biased dominance is also frequent in societies where female–female competition is intense, as in solitary or pair-living species where females are intolerant of each other, as well as in species where females face lower reproductive costs and are philopatric. Conversely, male-biased dominance is common in polygynous, dimorphic, terrestrial, and group-living species and often relies on physical superiority. In contrast, female empowerment hinges on alternative strategies, such as leveraging reproductive control. Our study highlights that male–female dominance relationships are highly variable and identifies the traits associated with the emergence of female- versus male-biased dominance in primate evolutionary history, which may also deepen our understanding of the origins of gender roles in early human societies.