Summary
Anthropic says Claude has supported work ranging from molecule design and laboratory testing to manufacturing documentation across six drug-development partnerships. The company describes the research as early and mostly preclinical.
Anthropic says it has worked with six organisations developing medicines to test how Claude can support tasks across drug development, from designing molecules to reviewing manufacturing records. In an update published on October 27, 2026, the company described early results from these partnerships and from its own drug-discovery work. It says most of the research remains preclinical, before medicines are tested in people.
The projects reflect a broad approach: applying AI not just to propose candidate molecules, but also to laboratory workflows, regulatory paperwork and decisions about which research programmes to pursue. Anthropic says the work has helped it identify areas for improvement in its models.
Claude’s role extends beyond molecule design
Drug development involves more than finding a molecule that interacts with a biological target. Researchers also need to refine a candidate’s properties, establish how it can be manufactured, prepare regulatory documents and plan clinical trials. Anthropic’s examples cover several of these stages, although the work described is at different levels of maturity.
At Generate Biomedicines, six scientists used Claude Code to propose shorter genetic instructions for a CAR T-cell therapy being adapted for solid tumours. The company says one design was about 70% shorter than the standard version. In laboratory tests, it enabled immune cells to make roughly 10 times as many receptors, killed tumour cells more efficiently and kept the immune cells active for longer.
Divergence Labs used Claude to develop antibody designs for a proposed food-allergy treatment. The approach aims to detach IgE, an antibody involved in allergic reactions, from immune cells. Anthropic says Claude generated roughly 35,000 designs in about a month. In early tests, Divergence found three designs that bound to IgE and 20 that could dislodge it; three of the latter appeared to do so more effectively than the standard medication in those tests. The next step, according to Anthropic, is to combine the two functions in a drug.
BigHat Biosciences used Claude alongside its laboratory data system to suggest changes to two antibody-fragment leads. Its laboratory made and tested about 1,010 variants over roughly 10 weeks. Anthropic reports that the best versions bound their targets more than 10 times as tightly as the originals, while one stayed in solution at a temperature 18°C higher and another clumped more than 100 times more slowly. The team also changed how it measured stability and clumping after reviewing the underlying data.
AI is also being applied to operational work
At Enveda, Claude reviewed manufacturing documentation and helped draft regulatory documents. Anthropic says that, in a test covering 1,061 pages, Claude flagged all 11 issues found by human reviewers as well as seven additional issues. The process reduced human review time by 89%, from more than 25 hours to less than three. Claude also drafted two regulatory documents; one took four hours to prepare, compared with about three months by hand, and Enveda’s standard review found one minor error in the other.
Anthropic’s other partnerships include work with The Column Group to establish Pleiades, a company developing tools and workflows for Claude to support multiple drug programmes. Its first Claude skill, designed to assess whether drug targets merit further study, has been evaluated on 45 target-disease pairs. Bristol Myers Squibb is using Claude to analyse an early-stage clinical-trial plan and estimate its prospects; the partners say they plan to expand this work in 2027.
Anthropic also says it acquired Mithril Medicine, a one-person company developing treatments for two rare forms of muscular dystrophy. Its founder used Claude for operational work as well as research coordination. A possible treatment for one form moved from an idea to promising results in mice in less than 18 months, according to Anthropic. The company expects a candidate could be ready to begin human trials as soon as late 2028 if it continues to hold up.
These reports illustrate where AI may contribute across a drug programme, including work that is not itself molecular discovery. But the reported laboratory and workflow results are early-stage measures, not evidence of clinical benefit. Anthropic characterises the overall effort as mostly preclinical, so whether the candidate medicines become safe and effective treatments will depend on subsequent research and trials.