OMatG-flash: An All-Atom Flow Map with Reinforce Adjoint Matching for Scalable Materials Discovery

The discovery of novel inorganic materials drives technological breakthroughs in critical fields such as computing and energy. Generative AI has promised to accelerate the discovery pipeline but state-of-the-art flow and diffusion models remain bottlenecked by the cost of proposing candidate materials. To address this, we introduce OMatG-flash, an all-atom flow map for inorganic crystal structure prediction (CSP) and de novo generation (DNG). OMatG-flash is a Pareto-optimal generator of candidate materials, producing high quality samples with an order of magnitude fewer inference steps and measured wall-clock time than existing flow and diffusion-based models while demonstrating raw benchmark performance competitive with the state-of-the-art. Furthermore, we extend the Reinforce Adjoint Matching post-training scheme to flow maps, achieving state-of-the-art performance at the CSP task. OMatG-flash showcases the potential of flow maps to accelerate generation of high-quality samples in the domain of materials science and demonstrates a step forward in sample throughput necessary for data-hungry materials science pipelines.