At Enveda, Claude drafted a filing in 4 hours vs 3 months

Claude wrote a first draft of a regulatory document for drug developer Enveda in four hours, a job that takes about three months by hand, Anthropic said in an October 27 update from Anthropic on six drug-development partnerships and its own drug programs. All of the figures in it are reported by Anthropic and its partners, and Anthropic says the work is still mostly preclinical: for the most part, the drug candidates involved have not yet been tested in humans.

The six partners are BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group, and the work runs from antibody design to manufacturing records and clinical trial planning. Anthropic argues that drug development has no single bottleneck, so Claude has to be involved across all of these stages, and says it has already prioritized specific model improvements based on what it has learned so far. The aim goes back to the Claude for Life Sciences launch in October 2025, when Anthropic said scientists typically used Claude for individual tasks such as statistics code or paper summaries and set the goal of supporting research “from early discovery through to translation and commercialization.”

Filings and trial plans

Before a clinical trial can start, a company needs records showing that each batch of the new drug was made to strict quality standards, plus an application to regulators. At Enveda, manufacturing chemists spent up to 75% of their time on paperwork. Enveda connected all of its manufacturing documents to Claude and had it flag possible issues. Across 1,061 pages of documentation, Claude caught all 11 issues human reviewers had spotted and flagged seven they had missed, and human review time fell by 89%, from over 25 hours to less than three.

Claude also wrote first drafts of two important regulatory documents: the four-hour one, and a second in which Enveda’s standard human review process found only one minor error.

More recently, Anthropic began working with Bristol Myers Squibb on clinical trial planning. BMS uses Claude to assess the trial plans of an early-stage drug in clinical trials and to estimate how likely those trials are to succeed, and the two companies plan to expand the work in 2027. “From a single prompt, Claude generated a robust initial analysis that helped inform our assessment of different development options, which would otherwise take days to complete,” said Michael Pourdehnad, BMS’s head of early clinical development for hematology, oncology and cell therapy. He added that BMS kept its highly sensitive data secure throughout.

Molecules in the lab

At BigHat Biosciences, which pairs AI with an automated lab, Claude was connected to the lab data system, Reccy, and proposed changes to two antibody fragments: one aimed at cancer cells, one meant to keep a drug in the body longer. The lab made and tested about 1,010 variants in about 10 weeks. Partway through, the team suspected the standard heat-stability and clumping metrics were misleading. Claude and an Anthropic researcher reanalyzed the raw data and proposed different metrics and a protocol change, BigHat’s scientists refined and ran it, and the main predictions held. Against the originals, the best versions bound more than 10x more tightly, the cancer-targeting fragment stayed in solution 18°C higher, and the other clumped more than 100x more slowly.

Divergence Labs wants a food-allergy drug that pulls IgE, the molecule that sets off the allergic reaction, off immune cells; standard medications only neutralize the IgE circulating in the blood. That needs a molecule with two arms, one binding IgE tightly and one prising it loose. Claude turned the goals into a design plan and then, largely autonomously, generated roughly 35,000 antibody designs in about a month. In early tests, three designs bound IgE and 20 dislodged it, and three of those 20 appeared to dislodge IgE more than the standard medication. Those are hits for each arm separately; building one drug that combines them is the next step. Anthropic made the same distinction in August about its own protein work: binders aren’t drugs.

Generate Biomedicines is trying to make CAR T-cell therapy, which has historically helped treat blood cancers, work on solid tumors by shortening one part of its genetic instructions to free room for others. Six Generate scientists used Claude Code to propose new designs for that part. One, about 70% shorter than the standard version, did better in lab tests: immune cells carrying it made roughly 10 times as many receptors, killed tumor cells more efficiently and stayed active longer.

Running a biotech on Claude

The Column Group is testing whether Claude can cut the time and cost of drug discovery enough that investors back riskier ideas over the previously validated ones they have favored. TCG is setting up Pleiades, a new company that will use Claude to run more drug programs, and more complex ones, than a small team usually can. Pleiades is still building tools and workflows; its first Claude skill, which assesses whether a drug target is worth pursuing, has been evaluated on 45 target-disease pairs.

Anthropic has also acquired Mithril Medicine, a one-person company developing treatments for two rare forms of muscular dystrophy. Its founder, Cody Tranbarger, used Claude to organize contracts, logistics and experiments with 11 external labs and delegated much of the rest of the company’s operations to it. Mithril took a possible treatment for one of the two forms from an idea to promising results in mice in less than 18 months. Tranbarger is continuing the research at Anthropic and plans to start several new programs. If his drug candidates keep holding up, Anthropic expects a drug to be ready to start human clinical trials as soon as late 2028.