Physiologically-based toxicokinetic ("pharmacokinetic") (PBPK or PBTK) modeling can be used as a tool to compare internal doses of acrylamide (AA) and its metabolite glycidamide (GA) in humans and rats. An earlier PBTK model for AA and GA in rats was refined and extended to humans based on new data. With adjustments to the previous parameters, excellent fits to a majority of the data for male Fisher 344 rats were obtained. Kinetic parameters for the human model were estimated based on fit to available human data for urinary metabolites of AA, and levels of hemoglobin adducts of AA and GA measured in studies in which human volunteers ingested known doses of AA. The simulations conducted with the rat and human models predicted that rats and humans ingesting comparable levels of AA (in mg/kg day) would have similar levels of GA in blood and tissues. This finding stands in contrast to the default approach that assumes a 3.2-fold increase in human risk due to pharmacokinetic differences between rats and humans. This model was used in a companion paper to estimate safe levels of ingested AA.