Drugs: Anthropic details his work with six partners and buys Mitthril, a one-person start-up against myopathy

CAR-T therapy shortened by 70%, 35,000 antibodies generated in one month, regulatory files written in four hours: Anthropic takes stock of Claude in the development of drugs.

Anthropic publishes an initial assessment of its life sciences team, which has been working for several months with six organizations: BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group. In it, the company defends a simple idea: drug bottlenecks are not limited to the design of molecules, they are spread across the entire chain, from the therapeutic target to clinical trials and regulatory dossiers.

Preclinical results

At Generate Biomedicines, Claude Code proposed a CAR-T therapy sequence that was about 70% shorter than the standard version; in the lab, it allowed immune cells to produce nearly ten times more receptors and kill tumor cells more efficiently. For Divergence Labs, which targets food allergies, the model generated some 35,000 antibodies in one month, three of which bind to IgE and twenty manage to detach it from cells. At BigHat, about 1,010 variants tested in ten weeks yielded molecules that bind more than ten times better to their target. Finally, at Enveda, Claude identified the 11 anomalies seen by human reviewers in 1,061 pages of manufacturing documentation, plus seven they had missed, and reduced review time by 89 percent.

Anthropic also announces the acquisition of Mithril Medicine, a one-person company founded by Cody Tranbarger, who has entrusted Claude with a large part of its logistics and exchanges with eleven laboratories. In less than 18 months, a treatment for a rare form of muscular dystrophy in children has gone from idea to encouraging results in mice; a human trial could begin at the end of 2028. The company points out that most of this work remains preclinical.