The coexistence of magmatic anhydrite and sulfide minerals in non-arc-related mafic magmas has only rarely been documented. Likewise the S isotope fractionation between sulfate and sulfide in mafic rocks has infrequently been measured. In the Kharaelakh intrusion associated with the world-famous Noril’sk ore district in Siberia coexisting magmatic anhydrite and sulfide minerals have been identified. Sulfur isotope compositions of the anhydrite–sulfide assemblages have been measured via both ion microprobe and conventional analyses to help elucidate the origin of the anhydrite–sulfide pairs. Magmatic anhydrite and chalcopyrite are characterized by δ 34S values between 18.8‰ and 22.8‰, and 9.3‰ and 13.2‰, respectfully. Coexisting anhydrite and chalcopyrite show Δ values that fall between 8.5‰ and 11.9‰. Anhydrite in the Kharaelakh intrusion is most readily explained by the assimilation of sulfate from country rocks; partial reduction to sulfide led to mixing between sulfate-derived sulfide and sulfide of mantle origin. The variable anhydrite and sulfide δ 34S values are a function of differing degrees of sulfate reduction, variable mixing of sulfate-derived and mantle sulfide, incomplete isotopic homogenization of the magma, and a lack of uniform attainment of isotopic equilibrium during subsolidus cooling. The δ 34S values of sulfide minerals have changed much less with cooling than have anhydrite values due in large part to the high sulfide/sulfate ratio. Variations in both sulfide and anhydrite δ 34S values indicate that isotopically distinct domains existed on a centimeter scale. Late stage hydrothermal anhydrite and pyrite also occur associated with Ca-rich hydrous alteration assemblages (e.g., thomsonite, prehnite, pectolite, epidote, xonotlite). δ 34S values of secondary hydrothermal anhydrite and pyrite determined by conventional analyses are in the same range as those of the magmatic minerals. Anhydrite–pyrite Δ values are in the 9.1–10.1‰ range, and are smaller than anticipated for the low temperatures indicated by the silicate alteration assemblages. The small Δ values are suggestive of either sulfate–sulfide isotopic disequilibrium or closure of the system to further exchange between ∼550 and 600 °C. Our results confirm the importance of the assimilation of externally derived sulfur in the generation of the elevated δ 34S values in the Kharaelakh intrusion, but highlight the sulfur isotopic variability that may occur in magmatic systems. In addition, our results confirm the need for more precise experimental determination of sulfate–sulfide sulfur isotope fractionation factors in high-T systems.