Electronic structure and direct dynamics calculations were used to study the potential energy surface and atomic-level dynamics for the OH- + CH3I reactions. The results are compared with crossed molecular beam, ion imaging experiments. The DFT/B97-1/ECP/d level of theory gives reaction energetics in good agreement with experiment and higher level calculations, and it was used for the direct dynamics simulations that were performed for reactant collision energies of 2.0, 1.0, 0.5, and 0.05 eV. Five different pathways are observed in the simulations, forming CH3OH + I-, CH 2I- + H2O, CH2 + I- + H2O, IOH- + CH3, and [CH3 - I - OH]-. The SN2 first pathway and the proton-transfer second pathway dominate the reaction dynamics. Though the reaction energetics favor the SN2 pathway, the proton-transfer pathway is more important except for the lowest collision energy. The relative ion yield determined from the simulations is in overall good agreement with experiment. Both the SN2 and proton-transfer pathways occur via direct rebound, direct stripping, and indirect mechanisms. Except for the highest collision energy, 70-90% of the indirect reaction for the SN2 pathway occurs via formation of the hydrogen-bonded OH- - -HCH2I prereaction complex. For the proton-transfer pathway the indirect reaction is more complex with the roundabout mechanism and formation of the OH- - -HCH2I and CH2I- - -HOH complexes contributing to the reaction. The majority of the SN2 reaction is direct at 2.0, 1.0, and 0.5 eV, dominated by stripping. At 0.05 eV the two direct mechanisms and the indirect mechanisms have nearly equal contributions. The majority of the proton-transfer pathway is direct stripping at 2.0, 1.0, and 0.5 eV, but the majority of the reaction is indirect at 0.05 eV. The product relative translational energy distributions are in good agreement with experiment for both the SN2 and proton-transfer pathways. For both, direct reaction preferentially transfers the product energy to relative translation, whereas transfer to product vibration is more important for the indirect reactions. For the proton-transfer reactions the velocity scattering angle distribution is peaked in the forward direction and in quite good agreement with experiment. However, for the S N2 reaction, the experimental scattering is isotropic in nature whereas forward scattering dominates the simulation distributions. The implication is that the simulations give too much stripping, which leads to forward scattering. The dynamics for the OH- + CH3I S N2 pathway are similar to those found previously for the F - + CH3I SN2 reaction.