This review/commentary addresses recent thermal and electromagnetic modeling studies that use image-based anthropomorphic human models to establish the local absorption of radiofrequency energy and the resulting increase in temperature in the body. The frequency range of present interest is from 100 MHz through the transition frequency (where the basic restrictions in exposure guidelines change from specific absorption rate to incident power density, which occurs at 3-10 GHz depending on the guideline). Several detailed thermal modeling studies are reviewed to compare a recently introduced dosimetric quantity, the heating factor, across different exposure conditions as related to the peak temperature rise in tissue that would be permitted by limits for local body exposure. The present review suggests that the heating factor is a robust quantity that is useful for normalizing exposures across different simulation models. Limitations include lack of information about the location in the body where peak absorption and peak temperature increases occur in each exposure scenario, which are needed for careful assessment of potential hazards. To the limited extent that comparisons are possible, the thermal model (which is based on Pennes' bioheat equation) agrees reasonably well with experimental data, notwithstanding the lack of theoretical rigor of the model and uncertainties in the model parameters. In particular, the blood flow parameter is both variable with physiological condition and largely determines the steady state temperature rise. We suggest an approach to define exposure limits above and below the transition frequency (the frequency at which the basic restriction changes from specific absorption rate to incident power density) to provide consistent levels of protection against thermal hazards. More research is needed to better validate the model and to improve thermal dosimetry in general. While modeling studies have considered the effects of variation in thickness of tissue layers, the effects of normal physiological variation in tissue blood flow have been relatively unexplored.