Abstract The prediction of cellular function from a genotype is a fundamental goal in biology. Constraint-based modelling methods systematize biochemical, genetic, and genomic knowledge into a mathematical framework enabling a mechanistic description of the genotype-phenotype relationship. The deployment of constraint-based approaches has evolved over roughly 30 years and since 2011, there has been a critical mass of studies combining them with high-throughput datasets and prospective experimentation. These studies have led to the validation of increasingly significant and relevant biological predictions. As reviewed herein, these recent successes have tangible implications in the fields of microbial evolution, interaction networks, genetic engineering, and drug discovery.