The assumption of gaseous species is not always valid in non-LTE plasmas. This requires the consideration of condensed phases in the determination of plasma composition. In this work, we compare two methods for calculating 2T multi-phase plasma composition: the Gibbs free energy minimization method and the entropy maximization method. The latter method is first reported in this work and has the same power as the former method for the calculation of 2T plasma composition taking into account condensed species. Based on these two methods, we present a case study of the multi-phase compositions of N2–Cu, CO2–Cu, and C2F4–Cu plasmas. It is found that the 2T entropy maximization method is less sensitive to the non-equilibrium degree than the 2T Gibbs free energy minimization method. This leads to the large difference between the composition predicted by the two methods, for example, in the phase transition temperature. The case study shows that due to the high electron temperature at large non-equilibrium degree, the ionic species (e.g. Cu+ and C+) can be generated directly from the condensed species (e.g CuF2, Cu2O, Cu, and C). Such ionization does not occur if the copper proportion is very low (e.g only 1%).