Since the introduction of solvent extraction as a separation method for metals, such solvating and acidic extractants have predominantly been used which are mono- or bifunctional compounds. They are able to form a five- or six-membered chelate ring, binding metal ions typically through one or two oxygen atoms, or through oxygen and nitrogen atoms. They can exhibit very useful selectivity, which is dependent on the localization of binding sites in various functional groups (acidic or neutral), chains or heterocyclic rings. However, a new approach in the search of new extractants is needed, if unusual selectivities are to be attained, e.g. within groups of chemically similar elements. Development of radius sensitive compounds represents one of the recent trends. The selectivity of macrocyclic ethers, mixed-heteroatom macrocyclic compounds and calixarenes is discussed in this review, with some emphasis laid on potential separation within groups of alkali and alkaline earth metals. Potential separation of lanthanides and actinides is also discussed, but less attention is paid to transition metals which can be separated by conventional extractants. Variables influencing the sensitivity are molecular configurations of the extractants nature of anions participating in the formation of extracted complexes, presence of mineral acids in the aqueous phase, nature of diluent and temperature. To separate elements with very similar ionic radii, extractants with less common, “soft” donor atoms can be applied. The separation of transplutonides (III) from lanthanides (III) is taken as an example, and the separation potential of acidic sulfur donor as well as neutral nitrogen donor extractants is discussed.