Dioxygen activation at nickel complexes is much less studied than for the biologically more relevant iron or copper systems but promises new reactivity patterns because of the distinct coordination chemistry of nickel. Here we report that a pyrazolate-based dinickel(II) dihydride complex [KL(Ni-H)] (1a) smoothly reacts with O via reductive H elimination to give the μ-1,2-peroxo dinickel(II) complex [KLNi(O)] (2a) and, after treatment with dibenzo[18]-crown-6, the separated ion pair [K(DB18C6)][LNi(O)] (2b); these are the first μ-1,2-peroxo dinickel intermediates to be characterized by X-ray diffraction. In 2a, the K is found side-on associated with the peroxo unit, revealing a pronounced weaking of the O-O bond: d(O-O) = 1.482(2) Å in 2a versus 1.465(2) in 2b; ν̃(O-O) = 720 cm in 2a versus 755 cm in 2b. Reaction of 1a (or 2a/2b) with an excess of O cleanly leads to [LNi(O)] (3), which was shown by X-ray crystallography ( d(O-O) = 1.326(2) Å), electron paramagnetic resonance and Raman spectroscopy (ν̃(O-O) = 1007 cm), magnetic measurements, and density functional theory calculations to feature two low-spin d nickel(II) ions and a genuine μ-1,2-superoxo ligand with the unpaired electron in the out-of-plane π* orbital. These μ-1,2-superoxo and μ-1,2-peroxo species, all containing the O-derived unit within the cleft of the dinickel(II) core, can be reversibly interconverted chemically and also electrochemically at very low potential ( E = -1.22 V vs Fc/Fc). Initial reactivity studies indicate that protonation of 2a, or reaction of 3 with TEMPO-H, ultimately gives the μ-hydroxo dinickel(II) complex [LNi(μ-OH)] (4). This work provides an entire new series of closely related and unusually rugged Ni/O intermediates, avoiding the use of unstable nickel(I) precursors but storing the redox equivalents for reductive O-binding in nickel(II) hydride bonds.