Metal−organic frameworks with open metal sites are promising materials for gas separations. Particularly, the M 2 (dobdc) (dobdc 4− = 2,5dioxidobenzenedicarboxylate, M 2+ = Co 2+ , Mn 2+ , Fe 2+ , ...) framework has been the Drosophila of this research field and has delivered groundbreaking results in terms of sorption selectivity. However, many studies focus on perfect two-component mixtures and use theoretical models, e.g., the ideal adsorbed solution theory, to calculate selectivities. Within this work, we shed light on the comparability of these selectivities with values obtained from propane/propene multicomponent measurements on the prototypical Co 2 (dobdc) framework, and we study the impact of impurities like water on the selectivity. Despite the expected capacity loss, the presence of water does not necessarily lead to a decreased selectivity. Density functional theory calculations of the binding energies prove that the water molecules adsorbed to the metal centers introduce new binding sites for the adsorbates.