Anthropic Unveils Claude Science to Turbocharge Scientific Research

Anthropic has officially launched Claude Science, a new AI built to support researchers in scientific discovery, drug development, and computational biology. The company unveiled the platform at a gathering of pharma companies, biotech startups, and academics. Anthropic says Claude Science is far more than a chatbot: it can autonomously run complex research tasks, write code, execute analyses, and document results in a fully reproducible way.

With this launch, Anthropic makes a major play in the fast-growing market for AI-driven science—long dominated by Google DeepMind.

Claude Science is a specialized research assistant, similar in spirit to Anthropic’s coding tool Claude Code, but purpose-built for scientific workflows.

Researchers can issue a high-level objective, and the system will break it down and execute the steps autonomously. Claude Science can:

That last point—reproducibility—is a key differentiator, according to Anthropic. Scientists must be able to trace exactly how an analysis or chart was produced, a core requirement in academic research.

While broadly applicable, Anthropic is initially targeting pharma and the life sciences.

During the demo, developer Alexander Tarashansky showed Claude Science autonomously surfacing potential treatments for phenylketonuria (PKU), a rare inherited metabolic disorder.

Anthropic also announced it will conduct its own research into treatments for rare and neglected diseases—both to contribute to medical progress and to build real-world experience with the platform.

Eric Kauderer-Abrams, head of Life Sciences at Anthropic, said the move aligns with the company’s mission to use AI for societal benefit.

The launch is strategically bold. For years, Google DeepMind has been the standard-bearer for AI in scientific research.

DeepMind achieved global fame with AlphaFold, the protein-structure prediction model that earned CEO Demis Hassabis and researcher John Jumper the Nobel Prize in Chemistry.

But with the rise of powerful AI agents and large language models, the landscape has shifted. Companies like Anthropic and OpenAI are building systems that can autonomously handle end-to-end research and programming tasks.

Anthropic is now pushing that shift directly into science.

Another signal: John Jumper recently left Google DeepMind to join Anthropic—bringing one of the world’s best-known AI scientists into the fold.

Anthropic argues that modern AI can now carry out meaningful research projects on its own.

Harvard physicist Matthew Schwartz previously wrote that Anthropic’s Opus 4.5 model performs at roughly the level of a second-year PhD student for research tasks. While subjective, it underscores how quickly AI is evolving from assistant to active research partner.

Claude Science advances that shift by not just generating answers, but automating entire scientific workflows.

The focus on drug development is both scientific and strategic.

Pharma companies spend billions of euros annually on R&D. AI can accelerate candidate discovery, streamline experiments, and significantly shorten timelines for bringing new medicines to market.

For Anthropic, that opens the door to a lucrative business opportunity. The company expects to post its first profitable quarter this year. Big pharma contracts could fuel that growth—especially as demand for generic AI chatbots starts to cool.

Claude Science signals the next phase of AI: beyond text and code, and deep into specialized knowledge work. Where AI once served mainly as a coding helper, Anthropic now positions it as a full-fledged research partner.

The competitive battleground is shifting from general chatbots to systems that empower entire professions. For pharma, universities, and research institutes, that could mean faster breakthroughs, lower costs, and a shorter path from basic science to new treatments.

With Claude Science, Anthropic isn’t just shipping a new product—it’s staking a clear claim: AI will power the next wave of scientific innovation.