Abstract The stability of Fe5O6 has been experimentally determined under pressure-temperature conditions relevant for the Earth’s deeper upper mantle down to the upper portion of the lower mantle (to 28 GPa). In addition, we investigated the incorporation of Mg into Fe5O6 and its systematics, which allows us to discuss the relevance of this phase for the mantle. Experiments were performed from 8–28 GPa and 900–1600 °C. Additional oxide phases may appear if the bulk composition does not maintain the Fe32+Fe23+O6 stoichiometry during the experiment, including coexisting Fe4O5 or Fe9O11. Unfortunately, the similarities in Raman spectra between several high-pressure Fe-oxide phases make this method unsuitable for distinguishing which phase is present in a given sample. The stability field for Fe5O6 extends from ~9 to at least 28 GPa but is truncated at lower temperatures by the assemblage Fe4O5 + wüstite. Refined thermodynamic properties for Fe5O6 are presented. The range of redox stability of Fe5O6 appears to be more limited than that of Fe4O5. Solid solution along the Fe5O6-Mg3Fe2O6 binary is quite limited, reaching a maximum Mg content of ~0.82 cations per formula unit (i.e., XMg3Fe2O6 ≈ 0.27) at 1400 °C and 10 GPa. The observed sharp decrease in molar volume of the O6-phase with Mg content could be a possible explanation for the limited range of solid solution. A phase diagram has been constructed for a composition of approximately Mg0.5Fe2.52+Fe23+O6 stoichiometry. This small amount of Mg causes a significant change in the relations between the O6-structured phase and the assemblage O5-structured phase + (Mg,Fe)O. Several experiments were performed to test whether the O6-phase can coexist with mantle silicates like wadsleyite and ringwoodite. In all cases, the run products contained (Mg,Fe)2Fe2O5 rather than the O6-phase, further underlining the limited ability of Fe5O6 to accommodate enough Mg to be stable in a mantle assemblage. The large stability field of Fe5O6 implies that this phase could likely occur in locally Fe-rich environments, like those sampled by some “deep” diamonds. However, the limited solubility of Mg in the O6-phase leads us to conclude that the O5-phase should be of much more relevance as an accessory phase in a peridotitic mantle assemblage.