Predicting others’ actions and inferring preferences from their choices is indispensable for successfully navigating social environments. Yet, the cognitive tools agents employ for prediction and decision may differ when involved in social interactions. When pursuing a goal individually, humans maximize utility by minimizing costs, while when engaged in joint actions utility maximization might not be the only heuristic in place. We investigate if human adults represent costs and rewards of joint vs. individual actions, and how do they decide whether to engage in a joint action. We test participants’ decisions when solving a task alone or together with a partner as a function of the cost of coordination. Our results show that human adults decide based on a preference for joint actions, despite engaging in coordination reduces their individual utility. We discuss a framework for decision-making which accounts for cognitive heuristics and preferences for joint actions characterizing agents’ cooperative behavior.