In life-cycle assessment (LCA) and comparative risk assessment, potential human exposure to toxic pollutants can be expressed as the population intake fraction (iF), which represents the fraction of the quantity emitted that enters the human population. To assess the influence of model differences in the calculation of the population iF ingestion and inhalation iFs of 365 substances emitted to air, freshwater, and soil were calculated with two commonly applied multimedia fate and exposure models, CalTOX and the uniform system for evaluation of substances adapted for life-cycle assessment (USES-LCA). The model comparison showed that differences in the iFs due to model choices were the lowest after emission to air and the highest after emission to soil. Inhalation iFs were more sensitive to model differences compared to ingestion iFs. The choice for a continental seawater compartment, vertical stratification of the soil compartment, rain and no-rain scenarios, and drinking water purification mainly clarify the relevant model differences found in population iFs. Furthermore, pH correction of chemical properties and aerosol-associated deposition on plants appeared to be important for dissociative organics and metals emitted to air, respectively. Finally, it was found that quantitative structure-activity relationship estimates for superhydrophobics may introduce considerable uncertainty in the calculation of population intake fractions.