A thin, active shell layer on core-shell columns provides high efficiency in HPLC at moderately high pressures. We revisited three models of mobile phase effects on retention for core-shell columns in mixed aqueous-organic mobile phases: linear solvent strength and Snyder-Soczewiński two-parameter models and a three-parameter model. For some compounds, two-parameter models show minor deviations from linearity due to neglect of possible minor retention in pure weak solvent, which is compensated for in the three-parameter model, which does not explicitly assume either the adsorption or the partition retention mechanism in normal- or reversed-phase systems. The model retention equation can be formulated as a function of solute retention factors of nonionic compounds in pure organic solvent and in pure water (or aqueous buffer) and of the volume fraction of an either aqueous or organic solvent component in a two-component mobile phase. With core-shell columns, the impervious solid core does not participate in the retention process. Hence, the thermodynamic retention factors, defined as the ratio of the mass of the analyte mass contained in the stationary phase to its mass in the mobile phase in the column, should not include the particle core volume. The values of the thermodynamic factors are lower than the retention factors determined using a convention including the inert core in the stationary phase. However, both conventions produce correct results if consistently used to predict the effects of changing mobile phase composition on retention. We compared three types of core-shell columns with C18-, phenyl-hexyl-, and biphenyl-bonded phases. The core-shell columns with phenyl-hexyl- and biphenyl-bonded ligands provided lower errors in two-parameter model predictions for alkylbenzenes, phenolic acids, and flavonoid compounds in comparison with C18-bonded ligands.