Claude Science, Anthropic’s flagship AI for life sciences

Anthropic announced Claude Science at an event for pharmaceutical executives, biotech founders, and researchers on Tuesday, launching a new flagship product that it says can autonomously carry out scientific work and is now available to all paid Claude subscribers. The product is aimed particularly at computational biology and drug development and will be used by Anthropic for its own research into drugs for rare, neglected diseases.

What is Claude Science?

Claude Science is a full-featured, standalone Claude product designed to support scientific research autonomously, with tool access geared toward genetics, chemistry, and protein biology. It parallels Claude Code and Claude Cowork in rank within Anthropic’s product lineup and is intended to write code, run that code on powerful computer clusters, and prioritize reproducibility so scientists can trace the source of any figure or result.

Claude Science follows Anthropic’s earlier release of plug-ins under the banner "Claude for Life Sciences," but unlike those plug-ins it is a distinct product. Alexander Tarashansky, who led its development, demonstrated how the system could autonomously identify new drug candidates for phenylketonuria during the event.

How does Claude Science differ from Claude Code and earlier tools?

Claude Science builds on the autonomous-work capabilities demonstrated in Claude Code while adding direct interfaces to scientific software, databases, and compute resources that many life-science projects require. Anthropic positions Claude Science alongside Claude Code and Claude Cowork rather than as a replacement, and it is meant to extend what scientists already find useful by handling both code creation and the often difficult task of running code on clusters.

Anthropic’s head of life sciences, Eric Kauderer-Abrams, framed the product as a mission-driven priority: "It represents how important this is to our mission that this is right up there with Claude Code and Claude Cowork as the next really significant product that we’re releasing." The company also emphasizes reproducibility as a core feature, so researchers can check and validate results.

What is the technical and competitive context?

Claude Science is built on Anthropic’s Opus model series and arrives as the company pushes further into scientific applications of LLMs that became capable of independent work in late 2025. The Harvard physicist Matthew Schwartz, writing about his work with Anthropic tools, estimated that the Opus 4.5 model "is about as capable of executing scientific projects as a second-year graduate student." Anthropic has also recently gained personnel momentum: John Jumper announced he is leaving DeepMind for Anthropic earlier this month.

For years Google DeepMind led many high-profile AI-for-science efforts, including AlphaFold, but the article notes that recent LLM progress has left DeepMind playing catch-up in some areas, especially coding. Anthropic says Claude Science can interface with various genetics, chemistry, and protein biology tools, making it useful for drug-hunting workflows.

Why it matters

Claude Science signals that Anthropic is prioritizing life sciences as a strategic product area and intends to compete for both scientific credibility and commercial revenue. The product launch coincides with Anthropic preparing to pursue its own drug-research projects for neglected diseases using Claude Science, and the company expects its first profitable quarter. A clearer commercial pathway—contracts with pharmaceutical companies—could help sustain profitability as Anthropic heads toward an IPO later this year.

What to watch

Watch for the outcomes of Anthropic’s internal drug-research projects using Claude Science and for any major contracts with pharmaceutical firms. Also track how Opus model iterations perform in real-world lab workflows and whether reproducibility claims hold up when outside labs attempt to replicate Claude Science–generated results.