An improved version of the FALCON fuel performance code is used to model the thermal-mechanical behaviour of the test rod in the CABRI high-burnup RIA test CIP0-1. First, the original FALCON code (version MOD01) is modified by introducing new routines yielding a better description of the initial fuel and clad geometry, which allows the modelling of highly oxidized claddings with non-uniform axial oxide thicknesses. By calculating the Hoop strain exerted by the oxide layer on the cladding, the correct initial relative position between inner and fuel clad outer surfaces is obtained. Application of the new algorithm to the REP-NA4 experiment and comparison with the previous code results shows that the new version yields better predictions of the final clad outer diameter profile. The new code version is then employed to predict the thermal-mechanical behaviour of the CIP0-1 test rod. The measured clad diametral deformation, clad elongation and coolant temperature are taken as figures of merit, the code calculations being compared against these. The agreement between code calculations and measurements is remarkable: (a) the final clad diameter profile is correctly predicted, provided that the transient oxide spalling is taken into account; (b) the clad elongation is satisfactorily reproduced if a bonded cladding and large axial friction factors are assumed; and (c) the coolant temperatures at different axial locations are generally overpredicted, but can be satisfactorily reproduced if the thermal inertia of the thermocouple is considered. [All rights reserved Elsevier]