Cooperation is the cornerstone of human societies, and its emergence is firmly linked to enhanced tolerance and egalitarianism. However, evidence of profuse cooperation in less tolerant and despotic societies challenges this predominant view. The overarching interdependency hypothesis may resolve the conundrum. It posits that group-level interdependencies, like strength in numbers in colonially nesting species or allomaternal care in cooperatively breeding species, promote indiscriminate cooperation through enhanced tolerance. Crucially, this hypothesis also predicts that dyadic interdependence like friendships, nepotistic biases, or coalitions, selectively enhance tolerance, fostering discriminate cooperation in despotic species. Species belonging to Macaca, which have a similar social organization, yet remarkable variation in tolerance, hierarchy steepness, nepotistic biases, and coalitionary tendencies, provide an opportunity for testing the interdependency hypothesis. In social group settings, we experimentally study cooperation, prosociality, and tolerance in six macaque species spanning a tolerance gradient. Our findings reveal high dyadic cooperation in despotic societies, yet this cooperation is restricted to a few partners. Dyadic prosociality, kinship, and tolerance positively predict cooperation. Further, our agent-based models demonstrate that despotic societies have fewer but more stable bonds and, thus, higher dyadic interdependencies than in egalitarian societies. Our results suggest that interdependencies facilitate the emergence and maintenance of cooperation.