The method of quasi-equilibrium directional crystallization was used for experimental modeling of the behavior of noble metals in the presence of Te during the fractional crystallization of Cu- and Ni-rich sulfide magma. The experimental melt contained (mol.%): Fe = 18.5, Ni = 19.1, Cu = 16.7, S = 44.1, and Pt = Pd = Rh = Ir = Ru = Ag = Au = Te = 0.2, i.e., is similar in composition to the massive pentlandite-bornite ores of platinum-copper-nickel deposits of the Noril’sk group. The crystallized sample consists of six primary zones differing in chemical and phase compositions. The main minerals crystallizing from the melt include the following sulfide phases: bornite solid solution (bnss), quaternary solid solution (tss), described earlier in the literature, and three phases (сfpn, cnpn, npn), which we attributed to pentlandite according to their chemical composition. The primary phases crystallized from the melt decay on cooling with the formation of secondary phases. The сfpn, cnpn, and tss phases decay completely, and the npn and bnss phases, partly. As a result, secondary zoning forms in the sample. Formation of drop-like inclusions of telluride melt was observed in the end zone of the ingot. The obtained data show that pentlandites and tss are the main high-temperature concentrators of PGE, with each of the macrophases showing specific PGE accumulation. Eight types of impurity phases have been detected. They form by different mechanisms: crystallization from sulfide melt of refractory compounds, deposition from telluride melt, and formation through complete or partial decay of primary macro- and microphases. A scheme of the zonal structure of the crystallized sample and the evolution of the phase composition during fractional crystallization has been constructed. It clearly demonstrates the intricate formation of primary and secondary major-component and impurity zonings and can be used to explain the nature of the zoned structure of massive PGE-bearing pentlandite–bornite orebodies.