Anthropic just bumped its AI for science effort from “plug-in” territory to full product status. On Tuesday, at an event for pharmaceutical executives, biotech founders, and researchers, the company announced Claude Science, a major new product intended to support scientific research in the same way that Claude Code supports software engineering. Like Claude Code, Claude Science can autonomously carry out meaningful work when given concise, high-level instructions, and it has access to tools that make it particularly useful for research in computational biology and drug development.

Along with launching and previewing Claude Science, Anthropic said it is now available to all paid Claude subscribers. That matters because this is not a small feature upgrade. Anthropic is elevating Claude Science alongside Claude Code and Claude Cowork, signaling that it wants the market to see scientific automation as a top-tier product category, not a side quest. At the same time, Anthropic said it will use Claude Science to pursue its own research into drugs for rare, neglected diseases.

The “why now” is partly strategic, partly operational, and partly a response to how quickly AI progress has been reshaping workflows. Over the past decade, Google DeepMind has been a leader in AI for science, highlighted by CEO Demis Hassabis and researcher John Jumper winning the Nobel Prize in chemistry for work on AlphaFold. DeepMind has also made contributions in meteorology and materials science. But the source frames a recent shift: in the past several months, the fast-advancing AI frontier seems to have left DeepMind “in the dust,” at least in the coding-centric application many large language model systems now dominate. The article also says DeepMind is “stuck playing catch-up” when it comes to coding.

Against that backdrop, Anthropic is trying to claim momentum. The source notes Anthropic CEO Dario Amodei is a PhD scientist, contrasting that with OpenAI CEO Sam Altman being described as a businessman. Whatever you think of executive bios, this still maps to a real buyer concern: if you are a scientist who needs AI assistance, credibility and domain fit matter. The company also positions Claude Science as a tool for researchers who are not expert software engineers. Today, scientific work often includes coding, but not all researchers can write, debug, and run code efficiently at scale.

This is where the product details start to matter for decision-makers. Anthropic had already released plug-ins under “Claude for Life Sciences” in October, allowing Claude to make use of scientific software and databases. But Claude Science is “a full-featured, standalone product,” not an add-on. That difference is important because standalone systems can change how work is organized: instead of scientists stitching together multiple steps themselves, an agent can coordinate the steps it needs. The source also says Claude Science prioritizes reproducibility, so scientists can trace back the source of any figure or result and check it for accuracy and validity. In science, reproducibility is not a nice-to-have. It is the difference between an output you can build on and a result you have to treat as suspicious.

The article adds two more practical workflow hooks. First, Claude Science writes code, but it can also help scientists run their code on powerful computer clusters, which the source says many scientists need but can find difficult to manage. Second, it is designed and marketed around molecular and cellular biology and drug development. That matters because those are heavy, tool-intensive domains, including workflows in genetics, chemistry, and protein biology, where interfaces to existing tools can reduce friction. During the Tuesday event, Alexander Tarashansky, who led the development of Claude Science, demonstrated how the system could autonomously identify new drug candidates for phenylketonuria, a rare genetic disease.

Finally, Anthropic’s incentive structure is not subtle. The source says it is not leaving this work solely to pharma companies and university labs represented at the event. Instead, Anthropic will pursue its own research into drug candidates for neglected diseases, both to help move science forward and to gain a clearer sense of how Claude Science works in the real world. There are obvious humanitarian reasons to prioritize drug development when building a general-purpose scientific research tool, and the source notes industry leaders often cite curing disease as a major potential upside of the technology. But the capital angle is also explicit in the reporting: pharmaceutical companies have far deeper pockets than academic researchers. The article further claims Anthropic says it is set to see its first profitable quarter, and that major new contracts with pharmaceutical companies could help keep it profitable as the tokenmaxxing craze dies down, especially as an IPO approaches later this year.

For executives and boards, the strategic stakes are straightforward. Claude Science is not just another model capability. It is a packaging decision, a go-to-market wedge, and a reproducibility-forward agent for computational biology, plus an internal pipeline push by the company itself. If the tool genuinely shortens the path from hypothesis to validated result, it can reshape how drug development teams allocate time between experimentation, coding, and verification. And if it lands in enterprise workflows early, Anthropic is positioning itself to own the category just as other labs and platforms figure out which parts of “AI for science” become procurement line items.