The dynamic development of experimental economics, in particular of experimental work in game theory, has contributed to the questioning of the basic assumptions of classical game theory: the rationality and self-interest assumption. In response, those assumptions are being modified and more realistic models of human behavior in situations which can be described as games are being developed. The article provides a short overview of the models that abandon the self-interest assumption. An account is given mostly of the distributional models (i.a. inequity aversion models, model of quasi-maximin preferences). The results of the experiments designed to test the rival distributional models of social preferences are also presented in the article, as well as of the experiments that assess to the role played by intention-based reciprocity in motivating human behavior.