The storage of granular solids in silos provides many interesting problems concerning pressures and flow. It is difficult to obtain repeatable and trustworthy results from either experimental studies or theoretical modelling. Comparisons of the best computational models with experiments are, at best, weak, and provide little assurance of the accuracy of any existing predictive model. The study described here was undertaken to explore the predictions of different models on a set of simplified exercise silo problems. For these problems, no experimental results exist, but simpler tests for truth can be used.This paper reports briefly on an international collaborative study into the predictive capacity of current discrete–element and finite–element calculations for the behaviour of granular solids in silos. The predictions of one research group, however eminent, are often not regarded as authoritative by others, so a commonly agreed theoretical solution of simple silo exercises, using different computational models from research groups around the world, is a valuable goal. Further, by setting the same unbiased exercise for both finite elements and discrete elements, a better understanding was sought of the relationships between the two methods and of the strengths of each method in practical silo modelling.The key findings are outlined here from three of the challenge problems: filling a silo, discharge of granular solid from a flat–bottomed silo; and discharge from a silo with a tapered hopper. Both computational methods display considerable shortcomings for these difficult exercises. Different research groups make widely different predictions, even when the problem statement is very detailed. There is much scope for further comparative studies to identify the reasons why different models based on comparable assumptions can produce such varied predictions.