Abstract The purpose of this work is to investigate the crystal growth parameters necessary or sufficient to obtain a crystal specifically of gem quality. We assume adequate chemistry is available. First, nucleation must occur with only a limited number of nuclei, otherwise too many crystals will be produced, and they will be too small to be faceted into a gem. Two growth mechanisms are readily documented for gems: Most commonly there is slow growth, driven by a spiral growth mechanism, leading to large single individuals. There are only a few examples of fast growth leading to gem-quality edifices: examples include “gota de aceite” Colombian emerald or the dendritic “pseudo cube” for gem diamonds. We have not documented the intermediate conditions between these two extremes in the Sunagawa diagram, which would correspond to 2D nucleation growth. The presence of inclusions is to be limited to desirable ones. Thus, in general, a good stability of the growth interface is the best guarantee of good clarity in the final gem. As for the interface, in general, growth conditions must be relatively stable over the period necessary to achieve growth. Perhaps surprisingly, it has become well documented that gem-quality near-colorless diamonds may have experienced quite a complex growth history. Therefore, the term stability has to be re-defined for each system producing a given gem. The length of time it takes to achieve crystallization of the gem has rarely been studied or estimated. Scientific evidence from experimental petrology and the growth of synthetic gems indicates that it does not take millions of years to grow a gem, but that this exercise may be achieved in a week to, arguably, a few years at the most. Available free space to grow does not appear always necessary, but it helps. Otherwise deformation, inclusions, and other negative effects may occur. Finally, no dramatic post-growth events, such as fracturing or etching, should affect the gem crystal.