Anthropic launches Claude Science tool for biology labs

Claude Science launches in beta for Anthropic paid users, aiming to automate biology and chemistry research tasks by combining common tools and 60+ databases.

Anthropic PBC said it will roll out the new software Tuesday, positioning it as a way to reduce repetitive work in scientific research while speeding up multistep analysis.

Automation aimed at biology and chemistry

The company said Claude Science can automate tasks across biology and chemistry, including work such as predicting protein structures. Anthropic framed the product as a practical layer on top of existing scientific workflows rather than a standalone lab platform.

Claude Science bundles tools scientists commonly use and connects them with more than 60 scientific databases, the company said at a Tuesday event. Anthropic said the goal is to reduce the need for researchers to jump across many separate sources when running an investigation.

Anthropic also emphasized a plain-language interface. The company said scientists can ask questions in everyday terms, automate multi-step tasks, and receive answers without writing numerous queries across different databases and tools.

Beta access and the product’s near-term limits

Anthropic said Claude Science will be offered as a beta to the company’s paid users. The company did not provide pricing changes, performance benchmarks, or validation results in the material provided.

That beta positioning matters because it signals the product is still being tested in real workflows. It also means that early user feedback is likely to shape how Claude Science handles sensitive scientific tasks, including how it cites sources across many databases and how it traces steps in multi-part analyses.

Anthropic steps into preclinical drug discovery

Separately, Anthropic said it has begun pursuing in-house preclinical drug discovery programs. That is a shift from selling tools to researchers toward running its own research efforts, at least at an early stage.

The company said it will focus on “areas that are outside the scope of what the traditional pharma and biotech landscape might consider attractive targets,” according to Eric Kauderer-Abrams, Anthropic’s head of life sciences, speaking at the event.

Anthropic did not specify which therapeutic areas it plans to pursue, how programs will be staffed, or whether it intends to partner with established drug developers. The announcement nonetheless places the company closer to the core economics of biotech, where preclinical work can be lengthy and expensive, and where validation standards are strict.

Part of a broader push for enterprise AI

Anthropic and rival OpenAI have spent the past year building AI products meant to expand beyond general chat into specialized professional use, including science and healthcare. The strategy is aimed at winning business customers and demonstrating that AI systems can deliver measurable productivity gains.

In Anthropic’s case, the company is attempting to translate its general-purpose Claude model into domain-focused tools that fit how professionals work. Claude Science follows that playbook by concentrating on workflows and data sources that scientists already rely on, rather than asking labs to rebuild processes from scratch.

Anthropic’s product cadence has also fed wider market anxiety about which types of services could be displaced by automation. The company pointed to earlier efforts to automate professional work, including the release in February of a tool for Claude Cowork intended to automate certain legal work such as contract review and legal briefings.

For life sciences, the immediate implication is sharper competition among AI vendors to become the interface layer for lab work, data retrieval, and multi-step analysis. If Claude Science proves reliable at stitching together results across dozens of databases, it could pressure incumbents that monetize access to specialized scientific information tools.

For Anthropic, the combination of a science product and in-house preclinical programs raises a strategic question: whether it plans to remain primarily a software provider, or evolve into a hybrid company that both sells tools and develops assets. The answer will shape how customers, partners, and regulators assess conflicts of interest and data governance.