Iambic unveiled Enchant v3, the next generation of its multimodal transformer model built to advance end-to-end drug discovery and development. Enchant v3 is trained using 41 billion parameters on more than 6,000 molecular properties across 16 biomedical data modalities, and it ingests diverse datasets, deploys a mixture-of-experts architecture, and tests multiple hypotheses in parallel to optimize discovery and development decisions. According to Co-Founder and CTO Fred Manby, Enchant v3 is an architectural leap over Enchant v2, with new capabilities to be deployed for internal and partner drug discovery and development efforts. Earlier versions demonstrated utility across Iambic’s pipeline, helping predict critical preclinical and clinical endpoints, and Enchant has shown robust scaling laws, with larger models exhibiting more powerful predictive capabilities. Enchant was built to ingest diverse data including public and proprietary data and modalities such as assays, text, images, multi-omics, biomolecular structures and sequences, and clinical trial data; leverage model scale; break through data walls separating preclinical drug discovery from clinical development; and convert model outputs into probability-guided decisions using uncertainty quantification. Highlights of v3 include 41 billion parameters, pretraining on 4.5 trillion tokens, 6,000+ molecular properties, 16 biomedical modalities, and predictable scaling laws.