Research interest in aprotic sodium-air (Na-O2) batteries is growing because of their considerably high theoretical specific energy and potentially better reversibility than lithium-air (Li-O2) batteries. While Li2O2 has been unequivocally identified as the major discharge product in Li-O2 batteries containing relatively stable electrolytes, a multitude of discharge products, including NaO2, Na2O2 and Na2O2·2H2O, have been reported for Na-O2 batteries and the corresponding cathodic electrochemistry remains incompletely understood. Herein, we provide molecular-level insights into the key mechanistic differences between Na-O2 and Li-O2 batteries based on gold electrodes in strictly dry, aprotic dimethyl sulfoxide electrolytes through a combination of in situ spectroelectrochemistry and density functional theory based modeling. While like Li-O2 batteries, the formation of oxygen reduction products (i.e., O2-, NaO2 and Na2O2) in Na-O2 batteries depends critically on the electrode potential, two factors lead to a better reversibility of Na-O2 electrochemistry, and are therefore highly beneficial to a viable rechargeable metal-air battery design: (i) only O2- and NaO2, and no Na2O2, form down to as low as ∼1.5 V vs. Na/Na+ during discharge; (ii) solid NaO2 is quite soluble and its formation and oxidation can proceed through micro-reversible EC (a chemical reaction of the product after the electron transfer) and CE (a chemical reaction preceding the electron transfer) processes, respectively, with O2- as the key intermediate.