The B3LYP density functional method has been validated for the di-Mn-substituted gamma-Keggin polyoxometalate (POM) anion, [(SiO4)MnIII2(OH)2W10O32]4-, and for the divacant lacunary silicodecatungastate, gamma-[(SiO4)W10O32]8-. This approach was shown to adequately describe the geometries of [(SiO4)MnIII2(OH)2W10O32]4- and gamma-[(SiO4)W10O32]8. Three different geometrical models, "full", "medium", and "small", for Mn2-gamma-Keggin have also been validated. It was shown that the medium [(SiO4)MnIII2(OH)2W6O24H8]4- model, as well as small [(SiO4)MnIII2(OH)2W4O18H10]2- model, preserves structural features of the full system, [(SiO4)MnIII2(OH)2W10O32]4-. However, the small model distorts the charge distribution at the "active site" of the system and should be used with caution. The same computational approach was employed to elucidate the structure of the di-Fe-substituted gamma-Keggin POM. The structure of the acidic (tetra-protonated form) of lacunary POM, gamma-[(SiO4)W10O32H4]4-, was shown to be gamma-[(SiO4)W10O28(OH)4]4- with four terminal hydroxo ligands, rather than gamma-[(SiO4)W10O30(H2O)2]4- with two aqua and two oxo(terminal) ligands as reported by Mizuno and co-workers (Science 2003, 300, 964). The observed and calculated asymmetry in the W-O(terminal) bond distances of gamma-[(SiO4)W10O32H4]4- is explained in terms of the existence of O1H1...O2H2 and O4H4...O3H3 hydrogen-bonding patterns in the gamma-[(SiO4)W10O28(OH)4]4- structure.