Komatiites require external sulfur from country rocks to generate immiscible sulfide liquid, which concentrates metals to form economic nickel sulfide deposits. Although signatures related to mass-independent fractionation of S isotopes (MIF-S, denoted as Δ33S) may identify external S sources, their values may not be directly indicative of the S reservoirs that were tapped during the ore-forming process, because of dilution by S exchange between assimilated sulfide xenomelt and komatiite silicate melt. To quantify this process and be confident that MIF-S can be effectively used to track S sources in magmatic systems, we investigated the effect of silicate melt-sulfide liquid batch equilibration, using the proxy of silicate/sulfide mass ratio, or R factor, on the resulting MIF-S signatures of pentlandite-rich ore from the Mount Keith MKD5 nickel sulfide deposit, Agnew-Wiluna greenstone belt, Western Australia. We carried out in situ multiple S isotope and platinum group element (PGE) analyses on pentlandite from a well-characterized drill core through the deposit. The variability in Pd tenor and MIF-S signature suggests that these are decoupled during batch equilibration and that the latter is not controlled by metal-derived R factor. Rather, the observed spread of MIF-S signatures implies that the sulfide xenomelt was initially heterogeneous and that chemical equilibration of S isotopes is incomplete as opposed to that of PGEs in a komatiite melt. Consequently, magmatic sulfides, which formed in the hottest, most dynamic, and likely fastest equilibrating magmatic systems on Earth, may still preserve their initial MIF-S isotope compositions, reflecting the range of crustal S reservoirs that were available upon komatiite emplacement.