A global evaluation is made of the four candidate models for IGRF 2000 (main field and secular variation) that were submitted to the IAGA Working Group V-8 by the deadline in February 1999. A description of the data used and the method of modelling is summarised for each candidate model. The models are then compared with one another and with recent data from observatories and magnetic surveys. It is concluded that none of the candidate main-field models is sufficiently good for an IGRF and that a new model should be derived using Orsted data.