Optimized geometries at the restricted Hartree-Fock (RHF), second-order Moller-Plesset perturbation (MP2) density functional theory (DFT), and configuration interaction with singles and doubles (CISD) levels are compared for the migratory insertion of NO into a Co-CH3 sigma-bond. Relative energies for these structures are examined along the reaction coordinate from the reactant (2) through the transition state (TS(a)) to the eta(1)-intermediate (3) at different levels of theory including higher levels of electron correlation such as quadratic configuration interaction with singles and doubles (QCISD), coupled cluster with singles and doubles (CCSD) and with perturbative corrections for triples (CCSD(T)), and CISD with size consistent corrections (CISD-(SCC)). DFT-B3LYP appears to give more reliable geometries in these first-row transition metal complexes than the RHF or MP2 approach. Although the MP2 optimized geometry of the product is in very good agreement with the experimental result, a near degeneracy problem affects the accuracy of the geometry optimization of the reactant, transition state, and eta(1)-intermediate. Because of this problem, the perturbation series (MP2, MP3, MP4) for the migratory insertion step fails to converge. Using higher level electron correlation methods such as CISD, CCSD, and QCISD are essential for energy calculations on this reaction. The CCSD//B3LYP method appears to yield the most reliable activation and reaction energies. This system is particularly sensitive to the theoretical method and would be useful as a model system for testing methods including electron correlation if better experimental values were available.