Extra-pair copulation (EPC) is widespread in socially monogamous species, but its evolutionary benefits remain controversial. Indirect genetic benefit hypotheses postulate that females engage in EPC to produce higher quality extra-pair offspring (EPO) than within-pair offspring (WPO). In contrast, the sexual conflict hypothesis posits that EPC is beneficial to males but not to females. Thus, under the sexual conflict hypothesis, EPO are predicted to be no fitter than WPO. We tested these two hypotheses with a 12 years dataset with complete life-history and pedigree information from an isolated island population of house sparrows (Passer domesticus). We compared fitness components of EPO and two types of WPO: (1) WPO from genetically polyandrous “unfaithful” mothers, and (2) WPO from genetically monogamous mothers. We found that all three groups of offspring had similar probabilities of hatching and nestling survival. Unexpectedly, EPO had the lowest probability of recruiting into the breeding population and the lowest lifetime reproductive output. Our results indicate that EPO incurred indirect genetic costs, rather than benefits, which is contrary to indirect benefit models. Importantly, the indirect costs we observed are also underappreciated in current sexual conflict models. Our results call for improved theoretical frameworks that incorporate indirect costs by extending current sexual conflict models.