Anthropic reports early, mostly preclinical drug development results with six partners and acquires Mithril Medicine

Anthropic published an update on October 27 on how its Claude models are being used in drug development, combining results from six industry partners with news that it has started drug programs of its own. The company describes the work as still at an early stage and mostly preclinical, meaning most of the candidates discussed have not been tested in people.

What the partners report

At BigHat Biosciences, Claude proposed changes to two antibody fragments; the lab made and tested about 1,010 variants over roughly ten weeks, and the best bound more than ten times more tightly than the starting molecules, according to Anthropic. Generate Biomedicines used Claude Code to redesign a component of a CAR T-cell therapy about 70% shorter than the standard version, which allowed roughly ten times as many receptors in lab tests. Divergence Labs generated around 35,000 antibody designs for a food allergy treatment in about a month, and three early candidates outperformed the standard medication at dislodging the allergy antibody IgE in lab assays.

Other uses are less about molecules. Enveda tested Claude on 1,061 pages of manufacturing records: it caught all 11 issues found by human reviewers plus seven they had missed, and cut review time by 89%. Bristol Myers Squibb uses Claude to assess clinical trial plans for an early-stage drug and plans to expand the work in 2027. The Column Group, a biotech investor, is setting up a company called Pleiades that will rely on Claude to run its programs. These figures are reported by Anthropic and its partners, not by independent studies.

Mithril and Anthropic's own programs

Anthropic also said it has acquired Mithril Medicine, a startup with a single employee, founder Cody Tranbarger, who has joined the company. Terms were not disclosed. Mithril works on two rare forms of muscular dystrophy that affect children; Claude helped coordinate contracts, logistics and experiments across 11 outside labs, and one candidate went from idea to promising results in mice in under 18 months. The Next Web reports that the lead program targets LAMA2-related congenital muscular dystrophy and notes that the underlying technical report has not been peer reviewed.

Anthropic says it plans to scale this work substantially and that a drug could be ready for human trials as soon as late 2028, provided the candidates keep holding up. That timeline depends on animal results holding up in further testing, so it is a conditional projection rather than a firm milestone.