Tokamak edge modelling codes are in widespread use to interpret and understand existing experiments, and to make predictions for future machines. Little direct benchmarking has been done between the codes, and the users of the codes have tended to concentrate on different experimental machines. An important validation step is to compare the codes for identical scenarios. In this paper, two of the major edge codes, SOLPS (B2.5-Eirene) and EDGE2DNIMBUS are benchmarked against each other. A set of boundary conditions, transport coefficients, etc. for a JET plasma were chosen, and the two codes were run on the same grid. Initially, large differences were seen in the resulting plasmas. These differences were traced to differing physics assumptions with respect to the parallel heat flux limits. Once these were switched off in SOLPS, or implemented and switched on in EDGE2D-NIMBUS, the remaining differences were small.