Anthropic launches Claude Science workbench for researchers
Claude Science combines research tools, specialist agents, scientific databases and computing resources in an auditable workspace.
Quick answer
What did Anthropic announce with Claude Science?
Anthropic introduced Claude Science, a research workbench that combines scientific tools, specialist agents, databases and computing resources. The beta is available on macOS and Linux for Claude Pro, Max, Team and Enterprise users, with features for reproducible analysis, literature work, figures, manuscripts and managed computing jobs.
Key takeaways
- Anthropic released Claude Science in beta for Claude Pro, Max, Team and Enterprise users on macOS and Linux.
- The workbench provides more than 60 curated skills and connectors for fields including genomics, single-cell research, proteomics, structural biology and cheminformatics.
- Claude Science records the code, computing environment, description and message history associated with generated figures.
- The app can work on a laptop, Linux machine or HPC login node and can use an existing HPC cluster or a Modal account for additional computing resources.
- Anthropic is offering credits to as many as 50 selected Claude Science AI for Science projects.
Anthropic has released Claude Science, an AI workbench that brings research tools, scientific data sources and computing resources into one environment. The app is available in beta on macOS and Linux for Claude Pro, Max, Team and Enterprise users.
Claude Science is designed to support literature analysis, multi-step research, figure creation and manuscript preparation. Anthropic says each output retains an auditable record of how it was produced, allowing researchers to examine and reproduce the work.
Agents and scientific resources
Users work with a generalist coordinating agent that has access to more than 60 curated skills and connectors. Anthropic says these resources are configured for genomics, single-cell research, proteomics, structural biology, cheminformatics and other fields.
The coordinating agent can create additional agents and work with specialist agents built by users. A reviewer agent examines citations and calculations, flags errors and makes corrections. Anthropic says the reviewer also checks for untraceable numbers and discrepancies between figures and their underlying code.
Claude Science is backed by more than 60 scientific databases. Anthropic identified resources including UniProt, PDB, Ensembl, Reactome, ClinVar, ChEMBL and GEO. The app also uses skills from Nvidia’s BioNeMo Agent Toolkit to connect with life-sciences models and libraries in BioNeMo, including Evo 2, Boltz-2 and OpenFold3.
Researchers can connect their own models, datasets and pipelines. Claude Science can save a pipeline as a reusable skill or access a laboratory’s preferred tool through a connector, with those additions available in later sessions.
Reproducible artifacts and managed compute
The app natively displays scientific artifacts such as 3D protein structures, genome-browser tracks and chemical structures. It generates figures and manuscripts alongside the code used to create them.
For a generated figure, Anthropic says Claude Science includes the exact code and environment, a plain-language description of the process and the complete message history. Researchers can request changes in plain language, such as removing gridlines or changing an axis to a logarithmic scale, and the agent modifies its code.
Claude Science can build computing environments and manage work on a laptop, a cluster or on-demand GPUs. It drafts a plan and requests permission before accessing new resources. Users can review or revoke decisions before the app writes and submits a computing job to an existing HPC cluster over SSH or to a Modal account.
Anthropic says the app runs on a laboratory’s own laptop, Linux machine or HPC login node. Large or sensitive datasets can remain on those systems, while only the context required for each analysis step is sent to Claude. Users can also fork a session to compare approaches without losing the original work.
Availability and project support
Team and Enterprise customers need an administrator to enable Claude Science. Anthropic also introduced discounted Team seats for active scientific laboratories at academic institutions and nonprofit research organizations.
The company said it will support as many as 50 Claude Science AI for Science projects with up to $30,000 in credits. Modal will provide up to $2,000 in computing resources for selected projects. Applications are open through July 15, 2026, with notifications scheduled for July 31 and projects running from September 1 through December 1.
Source: Anthropic’s Claude Science announcement, published June 30, 2026.
Frequently asked questions
- Who can use Claude Science?
- Anthropic says the beta is available to Claude Pro, Max, Team and Enterprise users. Team and Enterprise administrators must enable the app.
- Which operating systems does Claude Science support?
- The beta is available on macOS and Linux. Anthropic says it can run locally or on a remote machine through SSH or an HPC login node.
- How does Claude Science make work reproducible?
- Anthropic says generated figures include their code and environment, a plain-language account of their creation and the full message history. Outputs retain an auditable history.
- What scientific fields are supported?
- Claude Science is preconfigured for genomics, single-cell research, proteomics and cheminformatics. Anthropic also lists structural biology among the areas covered by its curated skills and connectors.