- Anthropic announced Claude Science, a major new product positioned to support scientific research the way Claude Code supports software engineering.
- It autonomously carries out research work from concise, high-level instructions, with tools for computational biology and drug development.
- The product is available now to all paid Claude subscribers.
- Anthropic will use Claude Science for its own research into drugs for rare, neglected diseases.
What Happened
Anthropic announced Claude Science, a major new product to support scientific research, according to MIT Technology Review. The company unveiled it at an event for pharmaceutical executives, biotech founders, and researchers, and elevated it to the same rank as Claude Code and Claude Cowork.
Like Claude Code, Claude Science can autonomously carry out meaningful work when given concise, high-level instructions.
Why It Matters
The launch extends Anthropic’s “autonomous agent for a profession” playbook from software to science. That playbook has worked before — Claude Code now authors about 4% of all public GitHub commits — and elevating Claude Science to flagship status signals Anthropic is betting AI-for-science is a category as large as coding, months after its $965 billion valuation.
Technical Details
Claude Science has access to tools that make it “particularly useful for research in computational biology and drug development,” and, like Claude Code, works from high-level instructions rather than step-by-step prompting. It is a full-featured, standalone product — a step up from Anthropic’s October release of “Claude for Life Sciences” plug-ins, which merely helped Claude use scientific software and databases. It is available immediately to all paid Claude subscribers.
Who’s Affected
Computational biologists, drug-development teams, and biotech founders are the direct audience. Anthropic itself is a user: it said it will apply Claude Science to its own research into drugs for rare, neglected diseases. Rivals pursuing AI-for-science — including OpenAI and Google DeepMind — now face a dedicated, flagship-tier competitor rather than a set of plug-ins.
What’s Next
The test is whether Claude Science produces validated scientific results, not just plausible ones — the bar in drug development is far higher than in code, where errors are cheap to catch. Anthropic’s own rare-disease work will be an early, watchable proof point for whether “autonomous research” holds up outside a demo.
