The ESM3 model, released by EvolutionaryScale, is a generative AI model for biology that can create novel proteins for use in drug discovery and materials science. Trained on a dataset of 2.78 billion proteins, ESM3 can reason over the sequence, structure, and function of proteins, enabling it to generate new proteins5. The model has been used to generate a new variant of green fluorescent protein (GFP). The full 98-billion-parameter model is available for non-commercial use through EvolutionaryScale's cloud Forge developer platform1.
ESM3, a generative AI model for biology, can create novel proteins for drug discovery, materials science, and carbon capture. It can also help scientists understand and engineer proteins, potentially leading to advancements in molecular biology, drug development, and synthetic biology. ESM3's ability to generate new fluorescent proteins could have implications for disease research and the development of new treatments.
EvolutionaryScale's $142 million seed funding round was led by ex-GitHub CEO Nat Friedman, Daniel Gross, and Lux Capital, with participation from Amazon and NVentures, Nvidia's corporate venture arm.