Image: Anthropic / anthropic.com Anthropic Launches Claude Science, a Research Workbench with 60+ Tools for Scientists
Claude Science is a purpose-built AI environment for computational research, connecting to genomics databases, proteomics tools, and structural biology resources in a single auditable workspace.
Anthropic launched Claude Science on June 30 — a research workbench aimed at scientists, built on top of Claude Opus 4.8 and available in beta to all paid subscribers on Pro, Max, Team, and Enterprise plans.
The product connects to more than 60 databases and tools covering genomics, proteomics, structural biology, and cheminformatics. When Claude Science generates a figure or analysis, it outputs the exact code and environment that produced it, a plain-language description of how it was made, and the full message history. The idea is reproducibility by default rather than by effort.
This is Anthropic’s most significant bet on a non-coding vertical. After Claude Code, which addressed software developers, Claude Science addresses research scientists. The pitch isn’t a new model — it runs on the same Opus 4.8 that subscribers already have access to — but a pre-configured environment that removes the setup friction of wiring together APIs, Python packages, and domain-specific databases.
What it connects to
The 60+ integrations span the databases researchers most commonly need:
- Genomics and sequence analysis tools
- Protein structure databases (structural biology)
- Cheminformatics resources for drug discovery work
- Proteomics analysis packages
The emphasis on auditable artifacts reflects a real problem in computational biology: it’s common for a published figure to be difficult or impossible to reproduce because the original environment wasn’t documented. Claude Science embeds the environment into the output.
Availability and the research grant program
Claude Science is available now in beta at no additional cost on Pro, Max, Team, and Enterprise plans.
Anthropic is also running an AI for Science grants program, offering up to $30,000 in credits to up to 50 research projects. Applications are open through July 15, with awards announced by July 31. Projects run September through December 2026.
The grant program targets the part of the market that’s most likely to generate public-facing results: academic and independent researchers who can publish findings. Getting a computational biology or drug discovery paper that relied on Claude Science into a journal creates a reference point that enterprise sales teams can use later.
Context
Claude Science follows a pattern Anthropic has been building toward. Claude Code brought the tool into software development workflows. Claude Science brings it into laboratory computing workflows. Both share the same approach: a domain-specific interface on top of a general-purpose model, with integrations that reduce the distance between “describe what you want” and getting a usable result.
Whether scientists adopt it depends on two things: whether the tool can handle the specificity of research workflows (a higher bar than general coding), and whether the audit trail holds up to the standards of peer review. Both are legitimately hard problems.
Source: Anthropic, TechCrunch