Anthropic Launches Claude Science to Bring AI-Powered Research into One Scientific Workspace
Advertisement
Claude Science is designed as a unified research environment where scientists can analyze published literature, process biological datasets, generate figures, write manuscripts, and execute large computational workflows.
Anthropic has launched Claude Science, a new research workbench designed for scientific research, with a strong focus on biology, biomedical science, and healthcare.
The workbench brings together literature analysis, computational tools, scientific databases, and high-performance computing to help researchers complete complex workflows faster while maintaining traceable and reproducible results.
The platform has integrated scientific databases, computational tools, coding environments, and scalable computing into a single workspace, allowing researchers to move from literature review to data analysis and manuscript preparation without switching between multiple applications.
The launch builds on Anthropic's broader push into healthcare and life sciences that began in late 2025.
The AI safety & research company is positioning the platform alongside growing AI adoption in biomedical research, an area that has gained significant momentum following advances such as DeepMind's AlphaFold for protein structure prediction and the increasing use of AI across genomics, drug discovery, and molecular biology.
Unified AI Workspace for Researchers
Claude Science is designed as a unified research environment where scientists can analyze published literature, process biological datasets, generate figures, write manuscripts, and execute large computational workflows.
Instead of functioning only as a chatbot, the platform coordinates specialized AI agents capable of handling genomics, proteomics, single-cell analysis, structural biology, cheminformatics, and other research domains.
The system comes preconfigured with more than 60 scientific databases and research tools, enabling researchers to query resources such as UniProt, PDB, Ensembl, Reactome, ClinVar, ChEMBL, and GEO through natural language.
Anthropic has also integrated NVIDIA's BioNeMo Agent Toolkit, allowing access to life sciences models including Evo 2, Boltz-2, and OpenFold3.
Laboratories can further customize the platform by connecting their own datasets, pipelines, and software through reusable skills and connectors.
A key feature of Claude Science is its emphasis on reproducibility. Every figure, chart, or manuscript generated by the platform is accompanied by the underlying code, software environment, workflow history, and a plain-language explanation of how the output was produced.
Researchers can request edits in natural language, while the system automatically updates the supporting code.
Reviewer agents continuously examine calculations, citations, and figures to identify inconsistencies and correct potential errors before results are finalized.
The platform also simplifies access to computing infrastructure. Scientists can submit computational jobs to local machines, laboratory servers, high-performance computing clusters over SSH, or cloud-based GPU resources through Modal.
Since analyses remain within the research environment, large datasets stay on institutional infrastructure, with only the necessary context exchanged during AI interactions.
Early Use Cases & Future Plans
Anthropic highlighted several early research applications from beta users.
Biotechnology company Manifold Bio has used Claude Science to prioritize tissue-targeting drug candidates by combining public biological data with internal research knowledge.
At the Allen Institute, neuroscientist Jérôme Lecoq developed a multi-agent workflow that reviews thousands of scientific papers, extracts evidence, generates comparative analyses, and produces long-form review articles with reviewer agents checking citation accuracy.
According to Anthropic, this workflow has reduced review-writing timelines from years to a fraction of the previous effort.
Stephen Francis, an epidemiologist at the UCSF Brain Tumor Center, has also reported using Claude Science to accelerate molecular epidemiology research on glioma while independently validating the platform's analytical results.
To encourage scientific adoption, Anthropic plans to support up to 50 AI for Science projects with as much as $30,000 in Claude credits, while compute partner Modal will provide up to $2,000 in computing resources for selected projects.
Anthropic also revealed that it has begun internal pre-clinical drug discovery programs using Claude Science to explore treatments for neglected diseases.
Stay tuned for more such updates on Digital Health News
