A coordinating agent, custom expert sub-agents, and a citation-checking reviewer give scientists one environment for computational research — running the same Claude models everyone already has.
2026-06-30 · by Moe Ameen
Anthropic introduced Claude Science on June 30, 2026, an AI workbench that gives scientists a single environment for computational research instead of bouncing between databases, pipelines, and tools. It launched in beta on macOS and Linux for Anthropic's Pro, Max, Team, and Enterprise plans, with a discounted Team plan for active scientific labs at academic institutions and nonprofit research organizations.
The product is built around an agent architecture. A generalist coordinating agent acts like a project lead, with access to 60+ curated databases (including UniProt, PDB, Ensembl, ClinVar, ChEMBL, and GEO) and prebuilt skills for fields such as genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. That agent can spin up sub-agents to split work or hand tasks to custom "expert" agents a user has built for their own research. A separate reviewer agent then checks citations and calculations, flagging and correcting errors before results reach a manuscript. The app natively renders scientific artifacts — protein structures, molecules, alignments, genomic tracks, PDFs — and preserves full lineage from raw data to publication-ready figures.
Anthropic was pointed about what Claude Science is not: not a new AI model and not a more capable model for biology. It runs the same Claude models already available to everyone, including Claude Opus 4.8, with no special access and no gating. The bet is that the workflow — coordination, reproducibility, and built-in fact-checking around existing models — is what wins scientists over, not a specialized model.
Anthropic pointed to early results from beta users: a neuroscientist at the Allen Institute described literature reviews that previously took up to two years now producing 100-page outputs with agent-checked citations, and a group at UCSF's Brain Tumor Center used it to speed up a comprehensive germline analysis of glioma, with results independently validated. Treat specific figures as Anthropic's reported claims rather than independent benchmarks.
If your lab or science brand ships research faster now, the new constraint is communication, not analysis — and that is the half Kompozy owns. Claude Science gets you a reviewed finding and a clean figure; Kompozy turns that into content an audience outside the lab will actually engage with. Drop the headline result and the figure into Kompozy as a source and the engine fans one finding into a science-explainer set: a persona or avatar video walking through the result, a Carousel that rebuilds the chart as branded slides, a Blog Article in plain language, an Email Newsletter for your list, and platform-native posts for LinkedIn, X, and beyond — all in your voice via a Persona Brief, then captioned, scheduled, and published across nine platforms from one queue. The science stays in Claude Science; the distribution stays in Kompozy.
There is also an immediate play in the launch itself. A workflow-over-model story like this is exactly the kind of timely topic your research-adjacent audience wants a take on this week. Put your point of view into Kompozy as a source and it can fan one angle into a blog post, a carousel explainer, short captioned clips, and platform-native posts, then schedule and publish them across your channels. Being early and clear on a story like this is how a single take becomes a week of content.
Claude Science is an AI research workbench from Anthropic, launched in beta on June 30, 2026. It gives scientists one environment for computational research — a coordinating agent with 60+ databases and prebuilt skills, custom expert sub-agents, a reviewer agent that checks citations and calculations, and reproducible outputs — on macOS and Linux for Pro, Max, Team, and Enterprise plans.
No. Anthropic was explicit that it is not a new model and not a specialized biology model. It runs the same Claude models already available to everyone, including Claude Opus 4.8, with no special access or gating. The bet is on the research workflow built around those models.
No. It is for doing and documenting research — analyses, figures, manuscript drafts — not for producing social posts, videos, or newsletters. To turn a finding into published, on-brand content across platforms you pair it with a content engine like Kompozy.