The predictions of a previously developed small cluster model for bound Cu ions and their interactions with CO and NO in Cu-exchanged zeolites [Schneider, W.F.; Hass, K.C,; Ramprasad, R; Adams, J.B. J. Phys. Chem. 1996, 100, 6032] are compared to those of a series of larger, more realistic cluster models. The hypothetical case of Cu-n+ (n = 0, 1, 2) coordinated to a single framework oxygen atom, described in the simplest model as CuOH2n+, is first examined in detail. The larger clusters considered include explicit Si or Al tetrahedral sites and are terminated by either H or OH. For nominal Cu(0) and Cu(I) oxidation states, the qualitative predictions of the original "water-ligand" model are found to be extremely robust, as are quantitative aspects of CO and NO binding. More pronounced differences between models are observed for a nominal Cu(II) oxidation state, due primarily to the tendency of the zeolitic portions of the larger clusters to reduce the Cu. The coordination of Cu-n+ (n = 1, 2) to a more realistic 4-fold coordination site model is similarly examined and compared to the results for Cu(OH2)(4)(n+). Excellent agreement is obtained between the water-ligand and cluster models for both Cu(I) and Cu(II) in this case. The results indicate that proper treatment of the local coordination environment of a Cu ion exchanged into a zeolite is of greater importance for describing chemistry on the Cu site than is use of a "realistic" model for an individual bridge oxygen site. The implications of these results for future modeling studies of Cu zeolites are discussed.