AI Advances Scientific Discovery
Anthropic Launches Claude Science AI Platform to Accelerate Life Sciences Research
Key Points
Claude Science can run locally on macOS or Linux or on a remote machine, and is designed to operate with per-specialist compute environments to manage workflows.
Team and Enterprise usage requires admin enablement, and Anthropic is offering discounted Team plans for research labs.
Outputs produced with Claude Science are auditable, carrying a history of how results were made to support validation and reproducibility.
The Claude Science announcement was unveiled at The Briefing: AI for Science, highlighting its role in accelerating research workflows.
Anthropic launched Claude Science on June 30, a desktop workbench for scientists that runs on macOS and Linux, according to 9to5Mac. The app enters beta alongside the existing Claude app and does not introduce a new AI model — instead, it acts as a workflow tool that connects researchers to more than 60 scientific databases, compute pipelines, and auditable outputs.
The debut happened at "The Briefing: AI for Science" in San Francisco, with Sanofi and the Broad Institute backing the platform from day one. Anthropic is offering up to $30,000 in API credits to 50 high-impact research projects and discounted Team plans for academic labs, signaling a clear push into life sciences revenue ahead of a possible IPO, The Tech Portal reports.
A Workbench, Not a Model, Targets Drug Discovery
Claude Science is best understood as an operating system for research. It plugs into databases like PubMed and Benchling, lets scientists generate figures through BioRender, and runs heavy compute jobs on local machines or remote servers via SSH. Ashu Singhal, president of Benchling, said the integration lets scientists move from "questions to decisions faster than ever."
For life sciences specifically, the platform taps NVIDIA's BioNeMo Agent Toolkit to call models like Evo 2 for genomics and Boltz-2 for binding affinity, Endpoints News reported. Every output carries a full history of how it was made — a feature Anthropic calls "provenance" — designed to support reproducibility and meet FDA-regulated research standards.
Claude Beats Human Experts on Protocol Quality Score
Anthropic shared a key benchmark at the launch. Claude Sonnet 4.5 scored 0.83 on a Protocol QA test. That beat the human expert baseline of 0.79 and the previous Sonnet 4 score of 0.74. Protocol QA measures how well an AI can answer questions about scientific lab procedures — a core task in drug development.
Real-world results back up the numbers. Stephen Francis of the UCSF Brain Tumor Center said his team completed germline workups in one-tenth the usual time using Claude Science. At Novo Nordisk, a task that previously took 10 weeks of documentation was reportedly cut to 10 minutes, according to Zamin.
Sanofi and Broad Institute Sign On as Launch Partners
Sanofi said Claude Science helps move research "from early discovery through to regulatory translation." The Broad Institute of MIT and Harvard called it a tool to help scientists pursue "ambitious questions in biology." Both endorsements give Anthropic credibility in a market where trust and data security matter enormously, Endpoints News noted.
Eric Kauderer-Abrams, Anthropic's head of Life Sciences R&D, said the goal is to give biologists the same speed boost that code generation gave software engineers. The platform supports local execution on macOS and Linux, which helps labs keep sensitive genomic files — often 100MB to 10GB in size — off third-party cloud servers.
Free Credits and Academic Plans Aim to Widen Access
Anthropic's AI for Science program offers up to $20,000 in API credits for individual researchers over six-month periods and up to $30,000 for 50 high-impact projects. The company also released 18 free courses under "Anthropic Academy" on June 29, including tracks on Agent Skills, to help researchers get started before the launch.
Still, some academic researchers worry that even discounted Team plans may be too costly for underfunded public university labs. That concern points to a broader risk: that Claude Science speeds up corporate R&D at companies like Sanofi while smaller labs fall further behind, The Tech Portal reported.
Sentiment
Mostly PositivePublishers
24
Articles
124
Reach
148
