Question: How well are genetic benefits hypotheses for extra-pair paternity supported by empirical evidence? Data incorporated: Almost all published studies testing for genetic benefits from 1980 onwards (121 papers, 55 species). Analysis methods: Collected key features and findings of each study in a database. Determined overall level of support for both good genes and compatible genes hypotheses. Conducted a formal meta-analysis on a subset of studies asking the following questions: (1) Do extra-pair mates of females have different phenotypes than their within-pair mates? (2) Do extra-pair offspring differ in viability from within-pair offspring? (3) Is there a correlation between the genetic similarity of a social pair and the incidence of extra-pair paternity? Results: Both the good genes and compatible genes hypotheses failed to be supported in more than half of the species studied. The meta-analysis shows that extra-pair males are on average larger and older than within-pair males, but not different in terms of secondary sexual traits, condition or relatedness to the female. No difference was found between extra-pair and within-pair young in survival to the next breeding season. We found no significant correlation between pair genetic similarity and extra-pair paternity. Conclusions: Genetic benefits are not strongly supported by available empirical data. New hypotheses are needed.