Claude Science Beta: An AI Workbench for Reproducible Research
Anthropic launched Claude Science in public beta -- a research environment, not a model. It runs analyses, queries 60+ scientific databases, and traces every step from data wrangling to publication, with code, environment, and conversation provenance attached to each artifact. It runs on your own infrastructure (laptop, HPC, GPU clusters) and submits jobs over SSH, Slurm, or Modal. Available now on macOS and Linux for Pro, Max, Team, and Enterprise plans.
Anthropic launched Claude Science in public beta: a dedicated research environment that is an app, not a model. It runs analyses, queries scientific databases, and traces every step from data wrangling to publication -- attaching the code, environment, and conversation history to each artifact it produces. It is available now for macOS and Linux on Pro, Max, Team, and Enterprise plans, and it runs on your own infrastructure so raw datasets and compute stay local.
The framing matters: this is a workbench wrapped around the model, not another chat surface. For how that fits Anthropic's broader push to turn Claude into a place where work happens rather than a box you paste into, see our coverage of Claude Opus 4.7 and the Claude design launch. For the team angle, our internal AI workspaces playbook covers how groups stand up shared, governed AI environments.
Key Takeaways
- It is an app, not a model. Claude Science is a research environment that orchestrates analysis, database queries, and compute -- it sits on top of Claude rather than being a new model release.
- Available now on macOS and Linux. Downloads cover macOS (Apple Silicon and Intel) and Linux. The feature is in beta and under active development.
- Plan gating. Offered on Pro, Max, Team, and Enterprise. On Team and Enterprise, an admin must enable it before members can use it.
- Your infrastructure, your data. It runs on a local laptop, Linux boxes, HPC login nodes, or GPU clusters; raw datasets and compute stay local.
- Reproducibility is the headline. Figures, tables, and notebooks carry attached code, environment, and conversation provenance, and a background reviewer flags bad citations and mismatched figures.
- Connects to 60+ scientific databases and submits jobs over SSH, Slurm-style HPC clusters, or a Modal account, scaling from one GPU to hundreds.
What Claude Science Is
Claude Science is, in Anthropic's words, "your research partner for rigorous science" that "runs analyses, searches databases, and traces every step from data wrangling to publication." The distinction worth internalizing up front: this is a public beta app, not a model. It is the environment Claude works inside, designed to connect to scientific databases, research tools, electronic lab notebooks, protein and structure models, HPC, and more.
That changes the unit of work. Instead of pasting context into a chat and copying results back out, the analysis, the data it touched, the compute it ran on, and the conversation that drove it live in one place -- and stay reconstructable afterward.
Availability and Plan Gating
Claude Science is available now for the platforms below, on Pro, Max, Team, and Enterprise plans.
| Platform | Status |
|---|---|
| macOS (Apple Silicon) | Available |
| macOS (Intel) | Available |
| Linux | Available |
On Team and Enterprise plans, an administrator must enable Claude Science before members can use it. Because it is in beta and under active development, admins should review the official documentation before a broad rollout.
Reproducibility and the Background Reviewer
The reproducibility story is the most concrete part of the pitch. Artifacts -- figures, tables, notebooks -- carry full provenance: the code, the environment, and the conversation history that produced them are attached to the artifact itself. Persistent Python and R kernels keep variables alive across sessions, so an analysis does not evaporate when you close the app.
On top of that sits a background reviewer that flags incorrect citations and mismatched figures, plus a plain-language annotation workflow for iterating on figures and drafting a manuscript alongside the analyses. The design goal is a chain you can trace from raw data to a publishable claim, with a check against untraceable or mismatched results along the way.
Compute, Databases, and Job Submission
Claude Science is compute-aware rather than compute-blind. It manages work across a local laptop, Linux boxes, HPC login nodes, and GPU clusters, scaling "from one GPU to hundreds" and writing batch scripts as needed. Job submission and management run over SSH on your own machine or HPC cluster, or through a Modal account.
It also connects to 60+ scientific databases so you can query them without learning each interface, and ships pre-configured domains for genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. Native viewing covers proteins, structures, molecules, alignments, genomic tracks, chemical structures, and PDFs.
When to Reach for Claude Science
| You want to... | Claude Science fit |
|---|---|
| Run reproducible analyses with attached provenance | Strong -- artifacts carry code, environment, and history |
| Keep raw datasets and compute on your own infrastructure | Strong -- runs locally and on your HPC/GPU |
| Submit and manage HPC or GPU jobs from one place | Strong -- SSH, HPC clusters, or Modal |
| Query many scientific databases without bespoke tooling | Strong -- 60+ databases connected |
| Roll out org-wide without admin review | Weak -- beta; Team/Enterprise admins must enable first |
Why This Matters
The interesting move here is positioning. Most AI-for-science effort goes into better models; Claude Science instead invests in the environment around the model -- provenance, compute management, database connectivity, and a reviewer that catches untraceable claims. That is the part of the research workflow that chat surfaces leave to the user, and it is where reproducibility actually breaks down.
The honest caveat is the one Anthropic states: this is beta, it may change, and on Team and Enterprise it is gated behind admin enablement. Treat it as an early but unusually concrete attempt to make AI-assisted research traceable end to end, not a finished platform.
FAQ
See the structured FAQ in the schema header for question-level details: what Claude Science is, which platforms and plans support it, where data and compute run, and how it supports reproducibility.
Sources
- Claude Science (beta): https://claude.com/science
- Anthropic (@claudeai) announcement: https://x.com/claudeai/status/2072002740830842899
Read next
Keep building the workspace playbookClaude Opus 4.7 Is Here: State-of-the-Art Coding, xhigh Effort, and a New Cyber Safeguards Tier
Anthropic launched Claude Opus 4.7 on April 16, 2026 -- a notable improvement on Opus 4.6 in advanced software engineering, with the same pricing, a new xhigh effort level, /ultrareview in Claude Code, higher-resolution vision, and the first deployment of cyber safeguards from the Mythos Preview track.
Ramp Glass Playbook: Internal AI Workspaces for Company-Wide AI Adoption
Ramp's internal 'Glass' platform reportedly pushed AI adoption to 99 percent by treating the harness -- not the model -- as the bottleneck. Every employee gets a fully configured workspace via SSO, a marketplace of 350+ shared skills, persistent memory from existing systems, and scheduled automations. Here is the playbook other companies can borrow.
Claude Design Launches: Anthropic Labs Turns Opus 4.7 Into a Prototype, Deck, and Wireframe Surface
Anthropic launched Claude Design on April 17, 2026 -- a research preview from Anthropic Labs that turns a prompt, uploaded image, or codebase into polished prototypes, pitch decks, and mockups. Powered by Claude Opus 4.7 vision, it learns your team's design system, exports to Canva, PDF, PPTX, or HTML, and packages finished designs for Claude Code handoff.
Frequently Asked Questions
What is Claude Science?
Claude Science is a public beta research environment from Anthropic -- an app, not a model. It runs analyses, queries scientific databases, and traces every step from data wrangling to publication. It connects to research tools, electronic lab notebooks, protein and structure models, and compute, and attaches code, environment, and conversation history to the artifacts it produces.
Which platforms and plans support Claude Science?
Claude Science is available now for macOS (Apple Silicon and Intel) and Linux. It is offered on Pro, Max, Team, and Enterprise plans. On Team and Enterprise plans, an admin must enable it before members can use it. The feature is in beta and under active development.
Where does my data and compute run?
Claude Science runs on your own infrastructure -- a local laptop, Linux boxes, HPC login nodes, or GPU clusters -- so raw datasets and compute stay local. It handles job submission and management over SSH on your own machine or HPC cluster, or through a Modal account, and scales from one GPU to hundreds.
How does Claude Science support reproducibility?
Artifacts such as figures, tables, and notebooks carry full provenance: the code, environment, and conversation history that produced them are attached. A background reviewer flags incorrect citations and mismatched figures, and persistent Python and R kernels keep variables alive across sessions so analyses stay reconstructable.