The separation between the B-3(1) and (1)A(1) states of Ni(O-2) is computed using density functional theory (DFT), internally contracted multireference configuration interaction with the multireference analogue of the Davidson correction (IC-MRCI+Q), and coupled-cluster with single and double (and perturbative) triple excitations (CCSD(T)) approaches. Although IC-MRCI+Q and DFT both predict a B-3(1) ground state, the CCSD(T) predicts a (1)A(1) ground state. We suggest that the true separation lies between the IC-MRCI+Q and CCSD(T) values, and that the ground state is likely (1)A(1), despite the IC-MRCI+Q result. We also show that Ni(O-2)(-) has a quartet ground state and, therefore, the (1)A(1) state of Ni(O-2) is probably not observed in the anion photoelectron spectra.