In late June, Anthropic announced the release of Claude Science, described as an “AI workbench for scientists” that is designed to expand the role of AI in scientific laboratories and accelerate research. The platform brings together multiple tools and databases within a single research environment, allowing scientists to conduct every stage of their work in one place.  

Claude Science enables researchers to analyze the scientific literature, refine figures and manuscripts, and perform other tasks throughout the research process. The platform also supports the creation of rich scientific artifacts, including 3D protein models, molecular visualizations, and genome browser views.  

Users can interact with a coordinating agent capable of performing more than 60 skills. The agent and its connectors are preconfigured with domain expertise in genomics, single-cell biology, proteomics, structural biology, cheminformatics, and related fields. Specialized agents can then retrieve and synthesize information from additional sources using their preferred query languages and methodologies. A separate reviewer agent is available to verify citations and calculations.  

Claude Science also integrates with NVIDIA’s BioNeMo life sciences agentic toolkit, as well as NVIDIA models and libraries such as Evo 2, Boltz-2, and OpenFold3, to support scientific discovery. Additional models, data sets, and pipelines can also be connected to the platform to provide greater flexibility and customization.  

All outputs, from figures and molecular structures to text, include information on how they were generated, along with a history of revisions to support transparency and reproducibility. Users can modify outputs using plain language without requiring coding expertise.  

Through its agentic workflow, the platform can perform multiple steps and complex analyses, including drafting plans and submitting computing jobs through an organization’s internal resources.  

Data sets remain within the laboratory’s infrastructure, helping to avoid memory limitations while addressing security concerns associated with proprietary information.  

Anthropic noted that Claude Science has already been used in scientific research for applications including single-cell RNA sequencing analysis, CRISPR screen design, and protein structure prediction.  

The system is available in beta for Claude Pro, Max, Team, and Enterprise users. It is accessible on macOS and Linux, including through remote access via SSH or high-performance computing login nodes.  

Drug Discovery 

During the launch event for Claude Science, Anthropic also announced that it is entering the drug discovery space, according to STAT. The company is launching an internal drug discovery program, with an initial focus on rare and “neglected diseases” (ie, conditions that have not traditionally been attractive targets for pharmaceutical companies). The specific disease targets have not yet been disclosed.  

Anthropic executives said this focus allows the company to avoid competing directly with its pharmaceutical customers while gaining insight into the models, products, and tools that the industry needs. The company will continue to provide AI tools to life sciences and pharmaceutical companies. 

Grants Program 

Since then, Anthropic has also announced a grant program for rare disease research through its AI for Science program, which provides free Anthropic credits to support high-impact scientific research.  When the program was introduced in May 2025, the initial cohort of selected scientific projects focused on biology and life sciences.  

“This rare disease grant program ties directly into our mission and work in beneficial deployments to extend the benefits of AI to areas that might not emerge naturally through market forces,” the company noted in its announcement.  

The company invited applications for research projects focused on rare genetic diseases and will award selected applicants $50,000 in Claude credits over 6 months. Grant recipients will also receive access to Claude Science.  

Half of the program is dedicated to basic rare disease research, with the remainder focused on clinical drug development. Example projects in the basic research track include developing evaluation tools to assess AI models’ performance on rare disease–related tasks, whereas examples for the drug development track include identifying measurable biomarkers and functional endpoints. According to Anthropic, the program is intended to help build a community of scientists using AI to advance the understanding of rare diseases.  

As part of the first track, Anthropic announced a partnership with the Monarch Initiative to use and contribute to the organization’s rare disease resources. The Monarch Initiative is an international consortium that integrates and interprets data to advance the diagnosis and treatment of rare diseases.  

Anthropic also suggested that AI could substantially accelerate drug development, clinical trials, and regulatory processes, and that its program would explore potential applications in these areas.  

Applications for the program are being accepted through August 2, 2026.

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.