Geothermal heat flux (GHF) is an important basal boundary condition for models of ice sheet dynamics, but is poorly constrained by conventional borehole-based estimates. The accuracy and uncertainties associated with geophysical proxy-based GHF models are contingent upon reliable models of heat production and thermal conductivity that are difficult to constrain. In this study, we examine (1) the statistical distribution of these thermal properties to gain insight into the Antarctic crust, (2) revaluate GHF estimates using these new constraints, and (3) identify areas where a lack of knowledge still hampers our ability to produce accurate models of GHF. Our approach centers on developing statistical models for the thermal properties based on global and regional distributions from global geochemical datasets. We then calibrate our heat production models of Antarctica using a combination of exposed Antarctic terranes, conjugate terranes and crustal tomography models of Antarctica. Regions where exposures and conjugate terranes are not accessible, we use a terrane model of Antarctica to reduce uncertainty. We then estimate GHF for different proxy-based temperature models using our thermal property calibrations. We find more diversity in GHF models derived from geophysical-based proxies when using a standardized crustal property model than predicted by the original. We suggest that there is still much to learn about the thermal state of Antarctica, which requires further improvements in predictors of thermal properties, especially their variation with depth.