A blind prediction exercise was organised on Indian Pressurised Heavy Water Reactor (PHWR) fuel to investigate the predictive capability of existing codes for their application at extended burnup and to identify areas of improvement. The blind problem for this exercise was based on a PHWR fuel bundle irradiated in Kakrapar Atomic Power Station-I (KAPS-I) up to about 15000 MWd/tU and subjected to detailed post-irradiation examination (PIE) in the hot cells facility at BARC. Eleven computer codes from seven countries participated in this exercise. The participants provided blind predictions of fuel temperature, fission gas release, internal gas pressure and other performance parameters for the fuel pins. The predictions were compared with the experimental PIE data which included fuel temperature derived from fuel restructuring, fission gas release measured by fuel pin puncturing, internal gas pressure in pin, cladding oxidation and fuel microstructural data. The details of the blind problem and an analysis of the results of blind predictions by the codes vis-à-vis measured data are provided in this paper.