June 27, 2015: We are pleased to announce version 1.3.3 with substantial improvements: Implement the relaxed clustering algorithm of PartitionFinder with -rcluster option to speed up analysis for data sets with many partitions (e.g., >100 partitions). Partition finding supports multicore now with linear speedup. Increased multicore performance of partition model analysis. Support direct translation of coding sequences into amino-acid (-st NT2AA option). Adjust per-partition state frequencies for partition model such that the log-likelihoods are fully comparable with RAxML. Print sequence identity scores along the tree (-wsi option). Fix slow convergence of partition model parameter estimation with many partitions. Fix memory allocation for large mixture models.