Anthropic has launched Claude Science, an AI workbench designed specifically for scientists, integrating the fragmented tools and databases researchers use every day into a single environment. Announced on June 30, 2026, the product is the company's most significant expansion yet of its push into the life sciences, and it arrives as the race to apply AI to drug discovery and biological research intensifies.
For researchers following the industry, the release is a notable step in how AI tools are reshaping scientific work — a trend we cover closely in our breaking AI news coverage. Where most AI assistants treat science as just another text-generation task, Claude Science is purpose-built for the lab.
One Environment for the Whole Workflow
Scientific research is notoriously tedious. Researchers must work across dozens of databases, each with its own schema, contend with file formats that require bespoke data pipelines, and juggle a roster of tools: PubMed, Jupyter, R, a cluster terminal, and more. Claude Science brings these fragmented tools into a single research environment where scientists can conduct every stage of their work, from analyzing literature to executing multi-step experiments.
Critically, every output carries an auditable history of how it was made. When the agent generates a figure, it includes the exact code and environment that produced it, a plain-language description of how it was created, and the full message history. That allows researchers to validate and reproduce results, even months later — a foundational requirement that traditional AI chatbots have routinely failed to meet.
Built for Biology From Day One
Claude Science is pre-configured for genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. Users interact with a generalist coordinating agent that has access to over 60 curated skills and connectors, backed by scientific databases including UniProt, PDB, Ensembl, Reactome, ClinVar, ChEMBL, and GEO. These agents can spin up other agents and engage with specialist agents that users create themselves.
The workbench uses NVIDIA's BioNeMo Agent Toolkit to connect natively to life-sciences models and libraries, including Evo 2, Boltz-2, and OpenFold3. It natively renders rich scientific artifacts — 3D protein structures, genome browser tracks, chemical structures — so researchers can inspect results visually rather than parsing raw output.
A Reviewer Agent That Checks the Work
One of the most distinctive features is a built-in reviewer agent that inspects outputs as a pipeline runs. It flags incorrect citations, untraceable numbers, and figures that don't match their underlying code, then self-corrects as it goes. That directly addresses one of the biggest fears scientists have about AI: hallucinated references and fabricated data undermining the integrity of research.
Researchers can ask Claude Science to make edits in plain language — removing gridlines, for example, or changing an axis to log scale — and the agent will rewrite its own code to comply.
Compute That Follows Your Lab
Large analyses, such as folding a protein or running a genomics pipeline over a massive dataset, normally force researchers to shift focus to setting up computing jobs and waiting on a cluster. Claude Science handles this automatically. It drafts a plan, asks before reaching new resources, and lets users review or revoke any decision before submitting the job to the computing resources a lab already uses — an HPC cluster over SSH, or a Modal account for compute on demand, scaling from a single GPU to hundreds as needed.
Because its agents work inside a running session that holds context in memory, even massive datasets only need to be loaded once. The workbench runs on a lab's own infrastructure — a laptop, Linux box, or HPC login node — so large or sensitive datasets never have to leave the systems they are already on.
Real Researchers, Real Results
Over the past several months, researchers tested Claude Science in beta for tasks like single-cell RNA sequencing analysis, CRISPR screen design, and protein structure prediction. Manifold Bio, which designs tissue-targeting medicines, used Claude Science to nominate targets for its latest experiments, assessing surface expression, trafficking, and safety across hundreds of candidates. Jérôme Lecoq, a neuroscientist at the Allen Institute, used it to build a multi-agent computational review template of roughly 20 custom skills that read through thousands of papers and construct long-form scientific reviews.
Claude Science is available now in beta for Claude Pro, Max, Team, and Enterprise users.
Stay Ahead of AI
As AI moves from chatbots into the lab, the implications for medicine and research are enormous. Keep up with every major release and breakthrough at AI Buzz Wire, your source for the latest AI developments.
Read more AI news →
