A deep understanding of the oxygen (O₂) reduction and evolution mechanisms is crucial for understanding metal–O₂ batteries. It has become evident that the instability of superoxide in the presence of lithium (Li) ions and sodium (Na) ions is the root cause for the poor reversibility and energy efficiency of Li–O₂ and Na–O₂ batteries. A straightforward yet elegant method is stabilizing superoxide with the larger potassium (K) ions. Superoxide-based K–O₂ batteries, invented by our group in 2013, are operated based on one-electron redox of O₂/potassium superoxide (KO₂) and have high energy efficiencies without any electrocatalysts. Nevertheless, limiting the anionic redox to O₂/superoxide affects the capacity output. Therefore, it is attractive to explore the possibility of beyond KO₂ in the K–O₂ batteries, especially if the use of catalysts can still be avoided. In this research, solid KO₂ was used as the condensed O₂ source and pre-dissolved in the dimethyl sulfoxide (DMSO)-based electrolyte. It is encouraging to observe two sets of reversible peaks during the three-electrode cyclic voltammetry scan under an argon atmosphere. One pair of peaks is attributed to the KO₂/potassium peroxide (K₂O₂) interconversion. Such redox has superb reversibility and a small overpotential of 239 mV in the absence of explicit electrocatalysts. Notably, it is further revealed that K₂O₂ reacts with gaseous O₂. Therefore, a gas-open system with an O₂ supply is unfavorable for realizing the reversible KO₂/K₂O₂ redox, and a closed cell system with a KO₂ supply as the starting active material is suggested instead.