Genetic evaluations have come a long way during the past decades, where the development and implementation of Best Linear Unbiased Prediction (BLUP) was undoubtedly the most notable achievement. The most important advances during the past 10 years were probably the direct use of test-day data in the BLUP model, ie. test-day models, the correction for heterogeneous within herd variances, multiple across country genetic evaluations (MACE), and the inclusion of more and more functional, and often difficult, traits in the evaluations. This paper will review the developments in test-day models, and the future of the genetic evaluations field, namely the inclusion of genomic information in the evaluations.