This article reviews the synthesis and properties of inorganic clusters that model the iron–sulfur centers found in the nitrogenase enzyme. These sites are of the Fe
4
S
4
, Fe
8
S
7
, and MoFe
7
S
9
stoichiometry. The enzyme [Fe
4
S
4
]
2+/1+
center is well modeled by synthetic clusters bearing thiolate ligands at iron. However, the [Fe
4
S
4
]
0
state is stabilized in the enzyme and has not been reproduced in the laboratory so far. Recent progress in synthetic models of the Fe
8
S
7
P‐cluster is reported. One of its known structures can be obtained by self‐assembly from basic constituents, or synthesized in a stepwise fashion from smaller clusters yielding topological analogues. The most challenging cluster to model is the FeMo‐cofactor. This MoFe
7
S
9
center is the site of dinitrogen reduction to ammonia. The cluster also exists with vanadium or iron in place of molybdenum in alternative nitrogenases. It was shown only recently that the FeMo‐cofactor is centered on a light atom that is probably a nitrogen atom. Structural modeling of this center is limited. Only its molybdenum or heterometal end has been accurately modeled. Functional chemistry of good structural models, those with sulfur ligands and metal–metal bonds, only concerns the reduction of already partially reduced diazo substrates such as diazene or hydrazine.