In the third workshop we will examine some of the most useful protein design models. Following a quick refresher we will explore "inpainting" as a super-tool for auto-completing one or several parts of any given design. We will next take a look at hallucination as a method to auto-generate de-novo proteins given a set of spatial restrictions. Finally we will review a recently published sequance design model (MPNN), which allows sequance optimization of any given protein structure.