Anthropic on October 27 described how its life sciences team has spent recent months embedding Claude in actual medicine programs rather than in isolated benchmarks. The company named six partners: Generate Biomedicines, Divergence Labs, BigHat Biosciences, Enveda, The Column Group and the pharmaceutical group Bristol Myers Squibb, and said it has started discovery work of its own. Each partner case is accompanied by a technical report or preprint, and Anthropic says it intends to expand the effort considerably.
The argument behind the work is that bottlenecks in drug development are spread across target selection, molecule design, manufacturing, regulatory paperwork and trial planning, so a model has to be useful in all of them. The case studies are chosen to show that range.
At Generate Biomedicines, six scientists used Claude Code to redesign part of the genetic instructions in a CAR T-cell therapy, with the goal of freeing space so the treatment could target solid tumors. One design was roughly 70% shorter than the standard one and, in lab tests, let immune cells produce about ten times more receptors and kill tumor cells more efficiently. Divergence Labs, which is pursuing a new approach to food allergies, had Claude turn its goals into a design plan and then generate about 35,000 antibody candidates in roughly a month, yielding early hits for both halves of a planned two-part drug.
BigHat Biosciences connected Claude to its lab data system and tested about 1,010 variants of two antibody fragments in some ten weeks. Anthropic says the best versions bound more than ten times more tightly, and that Claude and a researcher also spotted that standard stability metrics were misleading and proposed replacements. At Enveda, Claude reviewed 1,061 pages of manufacturing records, caught all 11 problems human reviewers had found plus seven they missed, and cut review time by 89%. It also drafted a regulatory document in four hours that would normally take about three months.
Bristol Myers Squibb is using Claude to assess plans for an early-stage clinical trial and estimate its odds of success, and the two companies plan to widen that work in 2027. With The Column Group, Anthropic is backing the creation of Pleiades, a new company meant to run more drug programs than a small team usually could by leaning on Claude.
The most unusual item is an acquisition. Anthropic has bought Mithril Medicine, a company run by a single founder, Cody Tranbarger, who used Claude to coordinate contracts, logistics and experiments with 11 outside labs while developing treatments for two rare childhood forms of muscular dystrophy. Anthropic says Mithril moved one candidate from idea to promising mouse results in under 18 months, and that a human trial could begin as early as late 2028 if results hold. Deal terms were not disclosed.
The limits are clear. Nearly all of the work is preclinical, the figures come from Anthropic and its partners, and lab results on molecules often fail to translate into approved medicines. The report also doubles as a sales pitch for Claude Science, Anthropic's research workbench, and for its Life Sciences Verification Program, which gives vetted organizations access to its strongest models under looser safeguards.