ABSTRACTThe benchmark dose (BMD) approach constitutes the most effective and pragmatic strategy for the derivation of a point of departure (PoD) for comparative potency analysis, risk assessment, and regulatory decision‐making. There is considerable controversy regarding the most appropriate benchmark response (BMR) for genotoxicity endpoints. This work employed the Slob (2017) Effect Size (ES) theory to define robust BMR values for the in vivo transgenic rodent (TGR) and Pig‐a mutagenicity endpoints. An extensive database of dose–response data was prepared and curated; BMD analyses were used to determine endpoint‐specific maxima (i.e., parameter c) and within‐group variance (i.e., var). Detailed analyses investigated the dependence of var on experimental factors such as tissue, administration route, treatment duration, and post‐exposure tissue sampling time. The overall lack of influence of these experimental factors on var permitted the determination of typical values for the endpoints investigated. Typical var for the TGR endpoint is 0.19; the value for the Pig‐a endpoint is 0.29. Endpoint‐specific var values were used to calculate endpoint‐specific BMR values; the values are 47% for TGR and 60% for Pig‐a. Endpoint‐specific BMR values were also calculated using the trimmed distribution of study‐specific standard deviation (SD) values for concurrent controls. Those analyses yielded endpoint‐specific BMR values for the TGR and Pig‐a endpoints of 33% and 58%, respectively. Considering the results obtained, and the in vivo genetic toxicity BMR values noted in the literature, we recommend a BMR of 50% for in vivo mutagenicity endpoints. The value can be employed to interpret mutagenicity dose–response data in a risk assessment context.