Many irradiation effects in Fe-Cr-Ni alloys such as radiation-induced segregation, radiation-enhanced diffusion, and void swelling are known to vary with bulk alloy composition. The development of microstructural and microchemical changes during irradiation and during post-irradiation annealing is determined by the rate of diffusion of point defects and alloying elements. To accurately predict the changes in grain boundary chemistry due to radiation-induced segregation and post-irradiation annealing, the composition dependence of diffusion parameters, such as the migration energy, must be known. A model has been developed which calculates migration energies using pair interaction energies, thereby accounting for the effect of composition on diffusivity. The advantages of this calculational method are that a single set of input parameters can be used for a wide range of bulk alloy compositions, and the effects of local order can easily be incorporated into the calculations. A description of the model is presented, and model calculations are compared to segregation measurements from seven different iron-chromium-nickel alloys, irradiated with protons to doses from 0.1 to 3.0 dpa at temperatures between 200°C and 600°C. Results show that segregation trends can be modeled using a single set of input parameters with the difference between model calculation and measurement being less than 5 at%, but usually less than 2 at%. Additionally, model predictions are compared to grain boundary composition measurements of neutron irradiated 304 stainless steel following annealing. For the limited annealing data available, model calculations correctly predict the magnitude and time scale for recovery of the grain boundary composition.