In this work, the solid–liquid equilibrium (SLE) curve for ten active pharmaceutical ingredients (APIs) with the polymer polyvinylpyrrolidone (PVP) K12 was purely predicted using the Conductor-like Screening Model for Real Solvents (COSMO-RS). In particular, two COSMO-RS-based strategies were followed (i.e., a traditional approach and an expedited approach), and their performances were compared. The veracity of the predicted SLE curves was assessed via a comparison with their respective SLE dataset that was obtained using the step-wise dissolution (S-WD) method. Overall, the COSMO-RS-based API–PVP K12 SLE curves were in satisfactory agreement with the S-WD-based data points. Of the twenty predicted SLE curves, only two were found to be in strong disagreement with the corresponding experimental values (both modeled using the expedited approach). Hence, it was recommended to use the traditional approach when predicting the API–polymer SLE curve. At the present moment, COSMO-RS may be an effective computational tool for the expeditious screening of API–polymer compatibility, particularly in the case of promising novel APIs, for which experimental datasets are likely limited or non-existent.