In this work we have summarized the available ab initio data addressing the interaction of carbon with vacancy defects in bcc Fe and performed additional calculations to extend the available dataset. Using an ab initio based parameterization, we apply object kinetic Monte Carlo (OKMC) simulations to model the process of isochronal annealing in bcc Fe doped with carbon to compare with experimental data. As a result of this work, we clarify that a binding energy of ~0.65 eV for a vacancy–carbon (V–C) pair fits the available experimental data best. It is found that the V 2–C complex is less stable than the V–C pair and its dissociation with activation energy of 0.55 + 0.49 eV also rationalizes a number of experimental data where the breakup of V–C complexes was assumed instead. From the summarized ab initio data, the subsequently obtained OKMC results and critical discussion, provided here, we suggest that the twofold interpretation of the V–C binding energy, which is believed to vary between 0.47 and 0.65 eV, depending on the ab initio approximation, should be removed.