There has been vigorous debate for several decades about whether the extreme enrichments of platinum group elements (PGEs) in some magmatic sulfide deposits could have resulted from simple equilibration of sulfide liquid with silicate melt. Key examples include the Ni-Cu-Pd mineralization in the Norilsk mining camp, the UG2 and Merensky reef Pt-Pd deposits in the Bushveld Complex, the Pd-rich J-M reef of the Stillwater Complex, and the Skaergaard Pd-Au mineralization. It was argued historically that the observed PGE tenors in these latter deposits are too high to be consistent with simple equilibration of sulfide and silicate melt. A commonly cited mechanism for increasing PGE tenor in magmatic sulfide is the upgrading of initially low tenor sulfide by allowing a small volume of sulfide to react with successive batches of fresh, previously undepleted silicate magma. Here we review several previous models for sulfide upgrading in light of recent changes in accepted values of the partition coefficients governing PGE exchange between sulfide and silicate, and we critically examine the physical scenarios implicit in each previous model. We show that, although sulfide upgrading may occur in natural settings such as fractional melting of the mantle, during the formation of sulfide accumulations from magmas it is unlikely to have effects that can be distinguished from simple one-stage batch equilibration. Even the most PGE-rich deposits currently known have compositions that can easily be accounted for by the simple one-stage batch process, with the possible exception of the Skaergaard Pd mineralization. It is generally not possible to use the measured composition of accumulations of magmatic sulfide to infer that sulfide upgrading has or has not occurred.