Accurate Co-C bond dissociation energies (BDEs) of large cobalamin derivatives in the gas phase and solution are crucial for understanding bond activation mechanisms in various enzymatic reactions. However, they are challenging for both experiment and theory as indicated by an obvious discrepancy between experimental and theoretical gas phase data for adenosylcobinamide. State-of-the-art dispersion-corrected DFT and LPNO-CCSD calculations are conducted for the Co-C BDEs of some neutral and positively charged cobalamin derivatives with adenosyl and methyl ligands and compared with available experimental gas phase and solution data to resolve the controversy. Our results from various levels of electronic structure theory are fully consistent with chemical and physical reasoning. We show undoubtedly that the Co-C bonds in complexes with the bulky adenosyl ligand are indirectly enhanced by many ligand-host noncovalent interactions and that the overall BDE are larger than those with the small methyl ligand in the gas phase. The additional intramolecular dispersion and hydrogen-bond interactions are significantly but not fully quenched in aqueous solution. The theoretical results including standard continuum solvation and dispersion corrections to DFT are in full accordance with experimental solution data. This is in agreement with several successful joined experimental/theoretical studies in recent years employing similar quantum chemical methodology. We see therefore no empirical basis for questioning the reliability of current dispersion corrections like D3 or VV10 to standard density functional approximations neither for these compounds nor for organometallic chemistry in general.