Given the intrinsic relation between phase behavior and performance in applications involving polymer mixtures, the thermodynamics of these systems has been the common focus of a wide body of experimental and theoretical studies. An elusive goal in this research field has been the ability to predict or simulate the degree of thermodynamic compatibility of two distinct homopolymers based on pure component properties alone. In an attempt to address this need, we present a simple model for the free energy of mixing of compressible polymer blends, based on a modification of the regular solution model. Its ability to qualitatively capture the phase behavior of weakly interacting polymer pairs using only the pure component properties of mass density, solubility parameter, and thermal expansion coefficient is demonstrated. To this end, a wide range of blend chemistries is considered, spanning purely dispersive polyolefin systems to those comprising more polar components, such as poly(methyl methacrylate), poly(epsilon -caprolactone), and polycarbonate. The thermodynamic and molecular origin of the observed phase behavior for each of the systems studied is also discussed.