All stories
Anthropic reports early Claude drug-discovery results

Anthropic reports early Claude drug-discovery results

Anthropic says Claude has produced measurable preclinical gains with six drug-development partners, including Bristol Myers Squibb, offering early evidence that AI can speed up more than the initial molecule-design work.

Anthropic reported results from several months of research partnerships with six organizations developing new medicines: BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group. The company has also started its own drug discovery programs, including work on rare childhood forms of muscular dystrophy, where it says Claude handles much of the scientific and operational work.

Anthropic argues that bottlenecks in drug development are spread across the whole pipeline, from target identification through regulatory filings and trial design, and are not limited to molecule generation. Among the examples it gave:

  • Generate Biomedicines: six scientists used Claude Code to redesign part of a CAR T-cell therapy. One design was about 70% shorter than the standard version. In lab tests it produced roughly 10x as many receptors, killed tumor cells more efficiently and kept immune cells active for longer.
  • Divergence Labs: working toward a food-allergy treatment, Claude generated roughly 35,000 antibody designs in about a month, largely on its own. Testing found three IgE binders and 20 designs that dislodged IgE, three of which appeared to outperform the standard medication.
  • BigHat Biosciences: connected to BigHat's lab data system, Claude proposed variants of two antibody fragments, and the lab tested about 1,010 of them in about 10 weeks. The best versions bound more than 10x more tightly than the originals. Claude also helped propose new stability metrics after the team questioned the standard ones.

All of these results are preclinical and self-reported, though partners have published technical reports. Anthropic says it has already prioritized model improvements based on the work and plans to scale it up substantially.

Sources