Anthropic Launches Claude Science and Its Own Drug-Discovery Program

Anthropic has launched Claude Science, an AI "workbench" built specifically for researchers — and, in a notable strategic turn, is spinning up its own in-house drug-discovery program. The product entered beta on June 30, 2026, and is available to Claude Pro, Max, Team, and Enterprise users, according to CNBC and MIT Technology Review.

Rather than a chatbot bolted onto a lab, Claude Science bundles more than 60 preconfigured tools, skills, and connectors into a single environment spanning genomics, proteomics, structural biology, and cheminformatics. It gives researchers flexible access to local, remote, and high-performance computing — and, crucially for science, produces auditable artifacts, with reviewer agents that check citations and recompute the numbers behind a claim.

Anthropic is now doing drug discovery itself

The bigger surprise is that Anthropic isn't just selling tools to pharma — it's standing up its own internal drug-discovery effort, initially aimed at neglected diseases. The company frames it as a learning exercise: by running real therapeutic programs in-house, it can work out which tools and models scientists actually need, then fold those lessons back into the product.

To seed adoption, Anthropic opened an AI for Science grant program — up to 50 projects will each receive as much as $30,000 in Claude credits. Applications are open through July 15, 2026, with awards announced by July 31.

The move plants Anthropic squarely in territory that Google DeepMind (via Isomorphic Labs and AlphaFold) and a wave of AI-bio startups already occupy — but with a different bet: a general-purpose, auditable research environment any lab can drive, rather than one specialized model. If it lands, Claude Science could become the everyday cockpit for computational research; and if the in-house drug program delivers, Anthropic will have proven the platform on its own molecules first.