Created in 2016 at Moscow State University’s Laboratory of Chemical Thermodynamics, the CpFit program is designed to approximate experimental data on the heat contents and isobaric heat capacities of individual substances using third-generation CALPHAD models that include weighted sums of Einstein functions. It supports approximating anomalies in heat capacity according to excess contributions. Support for robust regression and the ability of users to specify arbitrary anomalous contributions to heat capacity in the interpreted Lua programming language have also been added to the CpFit program. CpFit is a cross-platform free software distributed under the GNU GPL 2 license.