Anthropic Launches Claude Science, a Research Workbench for Drug Discovery, and Starts Its Own Neglected-Disease Drug Program

Overview

Anthropic has released Claude Science, an AI workbench for scientists, pushing its models deeper into laboratory research and pharmaceutical workflows. According to STAT, the company introduced the product on June 30, 2026, describing it as an application that optimizes its large language model for use in scientific laboratories and, especially, within the research operations of pharmaceutical companies. Alongside the launch, HealthCare Middle East & Africa reported that Anthropic has established an internal research programme of its own focused on developing treatments for neglected diseases.

What We Know

Claude Science is available in beta on macOS and Linux for Pro, Max, Team, and Enterprise plans, according to Anthropic. The company says the app ships with over 60 curated skills and connectors pre-configured for genomics, single-cell, proteomics, structural biology, and cheminformatics, and that a reviewer agent checks citations and calculations, flagging and correcting errors.

Those skills map to concrete laboratory tasks. Per HealthCare Middle East & Africa, researchers can design CRISPR screening experiments, analyse single-cell RNA sequencing data, and generate three-dimensional protein structures, chemistry drawings, and genome browser tracks, with integrations for common tools including PubMed, Jupyter, and R. pharmaphorum describes Claude Science as the first product specifically designed by Anthropic for use by life science researchers.

Anthropic positioned the release as a flagship effort. MIT Technology Review reported that the product is intended to support scientific research in the same way that Claude Code supports software engineering, and quoted Eric Kauderer-Abrams, Anthropic’s head of life sciences: “It represents how important this is to our mission that this is right up there with Claude Code and Claude Cowork as the next really significant product that we’re releasing. Our mission is to develop AI that serves humanity’s long-term well-being, and we believe that by far the greatest opportunity to do that is in the life sciences.” During the announcement, according to the same outlet, Alexander Tarashansky, who led the product’s development, demonstrated the system identifying drug candidates for phenylketonuria.

The company is also backing outside work. Anthropic said it will support up to 50 Claude Science AI for Science projects, providing up to $30,000 in credits, with applications open through July 15, 2026, and award notifications sent out by July 31. Named partners and connectors include NVIDIA’s BioNeMo Agent Toolkit, Modal, Manifold Bio, and the Allen Institute, per Anthropic. pharmaphorum reported that the workbench also connects to Basecamp Research’s EDEN, which it described as “the world’s largest biological dataset.”

Early institutional adoption preceded the launch. HealthCare Middle East & Africa noted that Bristol Myers Squibb signed a strategic agreement in May 2026 to deploy Claude across more than 30,000 employees. One beta tester, Stephen Francis, associate professor and epidemiologist at UCSF Brain Tumour Centre, told pharmaphorum: “After months of beta testing Claude Science, I’m convinced. This tool is going to accelerate scientific discovery in a big way.”

The launch extends a broader push into science. Anthropic recently hired Nobel laureate John Jumper from Google DeepMind, as previously reported.

What We Don’t Know

Anthropic has not detailed which specific neglected diseases its internal drug program will target, or on what timeline. The company’s public materials describe Claude Science as running on existing Claude models rather than a new biology-specific model, and independent, peer-reviewed validation of the workbench’s laboratory outputs has not yet been published. It also remains unclear how broadly pharmaceutical companies will integrate the tool into regulated discovery pipelines.

Analysis

The release reflects a strategic bet that the bottleneck in scientific AI is workflow friction rather than raw model capability. Anthropic CEO Dario Amodei framed the technology in broad terms, telling STAT that it would be “a general purpose technology that helps us to make sense of that complexity, in its full complexity, better.” By packaging domain skills, data connectors, and a citation-checking reviewer agent into a single environment, Anthropic is positioning Claude Science as an analog to Claude Code for the laboratory, and by launching its own neglected-disease program, it is signaling intent to be a drug discoverer, not only a tool vendor.