Automated model building, global search, TTE/mixture/Markov models, faster FOCEI and SAEM, a new covariate syntax, and benchmarks against NONMEM.
