Chai Discovery's AI Model Chai-2 Achieves Major Milestone in Rapid, Precise Antibody Design

SAN FRANCISCO--(BUSINESS WIRE)--Chai Discovery, a company at the forefront of using artificial intelligence to predict and reprogram interactions between biochemical molecules, has announced a significant breakthrough in de novo antibody design with its latest model, Chai-2. This advanced AI system achieves an impressive hit rate of nearly 20%, a substantial leap from traditional methods, which typically yield hit rates below 0.1%. Extensive laboratory validations of approximately 50 antibody targets demonstrated that nearly half produced validated hits, even when testing fewer than 20 designs per target. The designed antibodies exhibit desirable drug-like properties, such as nanomolar-range affinities, high specificity for their intended targets, and strong developability profiles, facilitating rapid translation into therapeutic applications. Additionally, Chai-2 has achieved equally impressive results with miniprotein targets, generating validated binders for 5 out of 5 test cases, with hit rates far surpassing previous state-of-the-art techniques. Mikael Dolsten, former Chief Scientific Officer of Pfizer for over 15 years, commented on the achievement: "I’ve been impressed by Chai Discovery’s ambitious goals since our early discussions—when such a breakthrough seemed distant—and it’s astounding to see it materialize so swiftly. Chai-2 has the potential to revolutionize the de novo design of medicines, significantly reducing development times." Antibodies play a crucial role in our immune systems and are vital for biomedical research and therapeutic discovery. Traditional methods of antibody development, such as immunization, directed evolution, and yeast-surface display, can be slow, expensive, and imprecise. Computational approaches have offered some speed and precision improvements, but they have historically struggled with low success rates and often required extensive and costly experimental screening to refine initial designs. Chai-2 stands out as the first AI system capable of reliably generating novel antibodies from scratch using only the target epitope as input. It requires no extensive lab screening or optimization rounds and supports a wide range of formats, including both nanobodies (VHHs) and the VH-VL format. Researchers can transition from in silico generation to lab validation in less than two weeks, drastically accelerating the development process. Joshua Meier, co-founder of Chai Discovery, highlighted the platform's capabilities: "Chai-2 is akin to Photoshop for proteins—it enables precise, rapid, and intuitive creation of biologic therapeutics, addressing targets that were once considered intractable. Our mission is to make the impossible possible, and this milestone marks a pivotal moment in scientific innovation." In a notable example, Chai-2 solved a complex antibody challenge that had previously required over $5 million in traditional R&D—in just a few hours. The solution was then validated in the lab within two weeks. With the biopharma industry facing pressures from patent expirations, cost-cutting measures, and the need to explore new therapeutic avenues, Chai's technology offers a transformative platform. It not only enhances productivity but also opens up entirely new areas of therapeutic research that were previously inaccessible. Chai Discovery has announced that it will provide early access to Chai-2 to select partners, fostering collaboration and further innovation in the field. For more detailed information, visit Chai Discovery's website or read their technical report. ABOUT CHAI DISCOVERY Chai Discovery is dedicated to transforming biology from a science into an engineering discipline by building cutting-edge artificial intelligence to predict and reprogram the interactions between biochemical molecules. The team consists of experts from leading research institutions and applied AI companies, including OpenAI, Meta FAIR, Stripe, and Google X. Chai Discovery is backed by prominent investors such as OpenAI, Thrive Capital, and Dimension.