For the first time, there was an extensive comparison between the plasma burn-through modelling codes presently available—DYON, SCENPLINT, and BKD0. The code benchmarking was carried out over three cases, adding more complexities in the modelling for each step. The first benchmarking was to revisit the mathematical models in the three codes. The simulation results were compared with the simplified settings, where ohmic burn-through of a pure hydrogen plasma was modelled with the ITER-relevant constant parameters. Without any reconciliation work, the three codes already predicted the same threshold prefill gas pressure for the hydrogen burn-through in ITER. However, the three codes produced different time-evolution of plasma parameters. The code developers identified differences in the mathematical models, which resulted in the different time-evolution. This enabled the codes to use the same mathematical models in the second benchmarking, which simulates an ohmic plasma burn-through discharge in a presently operating device. In the second benchmarking, the time-evolution of plasma parameters and impurities was simulated together with time-evolving experimental data from a JET discharge. It was found in this exercise that the time-evolving plasma volume and the electromagnetic modelling of eddy current in the passive structure could significantly change the simulation results, and thus their models need to be improved. For the third benchmarking, stand-alone simulations of the ECH modules in the three codes were compared. The calculated ECH absorption efficiency is identical in DYON and SCENPLINT, while it is about 20% higher in BKD0. Coupling the ECH modules, DYON and SCENPLINT simulated an ECH-assisted plasma burn-through discharge in KSTAR. In the early plasma burn-through phase, similar time-evolution of plasma parameters was reproduced. However, Te evolution starts to deviate from one another at about 100 ms. This is attributed to the sensitivity of ECH power absorption to the small initial difference in Te, which possibly resulted from the different atomic data in the two codes.