The design of single-molecule magnets (SMMs) based on the phenomenon of blocking of magnetization at low temperatures [1][2][3][4][5][6] has become a hot area of research due to the potential applications of such compounds in new storage and information-processing technologies. [7][8][9][10] Most of the SMMs synthesized so far are polynuclear transition metal complexes at their strong exchange limit, [11] where the exchange splitting is much larger than the zero-field splitting on individual metal sites. An SMM effect is obtained in the case of axial zero-field splitting, DS z 2 , where S z is the projection of the total spin S on the symmetry axis of the cluster, with D < 0. In order to have a large barrier for reversal of magnetization, j D j S 2 , these complexes should possess a large S and large magnetic anisotropy projected on the ground exchange multiplet (D).At the opposite limit of weak exchange coupling the zerofield splitting on metal sites is much larger than the exchange splitting in the complex.[11] This is always the case for complexes containing lanthanides and actinides and is often so for transition metal complexes containing cobalt ions or second-and third-row transition metal ions. The high anisotropy on lanthanide ions has prompted investigations of the effect of their incorporation on the SMM performance [12][13][14][15][16][17] as both the ionic anisotropy and exchange interaction in these complexes can contribute to the height of the barrier of reversal of magnetization. The first mechanism alone is responsible for the SMM effect in the mononuclear bis(phthalocyaninato)holmium anion. [13] A pure exchange contribution to the barrier has been predicted for some complexes with axial symmetry involving heptacyanomolybdenum(III).[18] Generally, however, the origin of SMM behavior of complexes at the weak exchange limit is difficult to elucidate. = 1,1,1-trifluoro-7-hydroxy-4-methyl-5-azahept-3-en-2-one), [20] have recently been synthesized and investigated. All these compounds show slow relaxation of magnetization. Compound 3 is characterized by a relatively high anisotropy barrier of 25 K, whereas 1 and 2 show a vanishingly small susceptibility at low temperature, which was completely unexpected for systems containing an odd number of electrons. Both these latter compounds show similar magnetic properties despite the presence of very different magnetic networks, [19] which allows intermolecular antiferromagnetic exchange interactions to be ruled out as a reason for this vanishingly small susceptibility at low temperature. In view of such an unprecedented situation, we have investigated the local anisotropy of the dysprosium sites in these complexes by high-level ab initio calculations and have simulated, on their basis, the lowest exchange states. This procedure has allowed an unambiguous determination of the nature of the ground state in these complexes.The main structural difference between the triangular units in 1 and 2 (Figure 1) is the Dy(3) site, which is coordinated by one chloride ion...