Results of 2D disruption modelling for validation of benchmark ITER scenarios using two established codes—DINA and TSC, are compared. Although the simulation models employed in those two codes ought to be equivalent in the resistive time scale, quite different defining equations and formulations are adopted in their approaches. Moreover there are considerable differences in the implemented model of solid conducting structures placed on the periphery of the plasma such as the vacuum vessel and blanket modules. Thus it has long been unanswered whether the one of the two codes is really able to reproduce the other's results correctly, since a large number of code-wise differences render the comparison task exceedingly complicated. In this paper, it is demonstrated that after the simulations are set up accounting for the model differences, a reasonably good agreement is generally obtained, corroborating the correctness of the code results. When the halo current generation and its poloidal path in the first wall are included, however, the situation is more complicated. Because of the surface averaged treatment of the magnetic field (current density) diffusion equation, DINA can only approximately handle the poloidal electric currents in the first wall that cross the field lines. Validation is carried out for DINA simulations of the halo current generation by comparing with TSC simulations, where the treatment of halo current dynamics is more justifiable. The specific details of each code, affecting the consequence in ITER disruption prediction, are highlighted and discussed.