We have carried out a theoretical study on the hydrogen and chlorine abstraction reactions by silyl (H3Si .) and trichlorosilyl (Cl3Si .) radical from the three chloromethanes ClCH3, Cl2CH(2), and Cl3CH. The results of traditional ab initio methods CHF, MP2, and MP4) have been compared with those obtained with a DFT approach using the hybrid B3LYP and the pure BLYP functional. At all computational levels, we have found that the chlorine abstraction is highly favored with respect to the hydrogen abstraction as experimentally found. Furthermore, not only the Hartree-Fock method, as one can expect, but also correlated methods such as MP2 and MP4 largely overestimate the activation barriers. Only the DFT approach is able to provide results in good agreement with the experiment; in particular, the best performance has been obtained with the hybrid functional B3LYP which seems to be particularly suitable for investigating the reactivity of these types of radicals. A simple diabatic model based upon valence bond theory has been used to rationalize the reactivity pattern shown by these radical reactions.