Anthropic's Claude Science AI workbench now integrates with NVIDIA's BioNeMo Agent Toolkit to let scientists run complex
NVIDIA has spent over a decade building a full GPU-accelerated computing stack for life sciences, including hardware, frameworks, libraries, models, microservices and domain-specific tools that help researchers run sophisticated workflows and iterate faster.
This week, Anthropic announced Claude Science, an AI workbench designed for science research that enables scientists to converse with agents in natural language to run their work end to end. Claude Science integrates directly with NVIDIA's BioNeMo Agent Toolkit as a resource that scientists can access within their workflow. The toolkit packages NVIDIA-accelerated capabilities as callable skills, allowing Claude Science to select the appropriate tool, prepare valid inputs and execute the workflow while connecting to NVIDIA compute resources deployed anywhere. This integration brings NVIDIA's accelerated models, libraries and NIM microservices directly into the environment where the research happens.
The largest pharmaceutical companies worldwide rely on NVIDIA technologies to advance AI-enabled research across drug discovery, genomics, medical imaging, molecular design and protein engineering. Eighteen of the top twenty pharmaceutical companies currently use NVIDIA BioNeMo, demonstrating its critical role across the ecosystem.
Claude Science enables scientists to move their research from intent into action using natural language, without manually configuring models, endpoints or software environments. NVIDIA BioNeMo Agent Toolkit extends this capability by providing access to accelerated workflows and models including Evo 2, Boltz-2 and OpenFold3, so analyses that benefit from acceleration run significantly faster.
A scientist begins by describing a research task in natural language, such as analyzing a genomic sequence, predicting a protein structure or designing a potential binder. Claude Science interprets the request and orchestrates the work through preconfigured domain-specialized agents that understand established workflows across genomics, proteomics, single-cell analysis, cheminformatics and clinical research. BioNeMo Agent Toolkit provides these agents with the context needed to connect each step with an appropriate NVIDIA scientific capability. Each skill includes information about its purpose and required inputs, helping agents prepare and execute the workflow and return outputs for review.
The result is an iterative loop between scientific reasoning and accelerated computational work. Scientists can inspect outputs, refine their questions and determine the next step while staying focused on the science.
A practical example is generating improved inhibitors of common cancer targets. In this workflow, a scientist starts with a known cancer-causing antigen mutation and asks Claude to design numerous potential inhibitors. With Claude Science integrated with BioNeMo Agent Toolkit and NVIDIA NIM microservices, the system accelerates high-throughput inhibitor prediction, optimization and validation.
Autonomous AI scientist agents don't reason in isolation. They may need to fingerprint a library of compounds, cluster promising hits, generate conformers for top candidates, analyze genomic context and compare perturbation responses before recommending the next experiment. Each step relies on a scientific tool, and the agent can only work as fast as those tools run. NVIDIA BioNeMo Agent Toolkit gives scientific agents the accelerated tools they need to operate at the speed of science.
The toolkit is open and framework-agnostic, allowing the same scientific skills to work across different agent frameworks and research platforms. The toolkit and its skills are available now through NVIDIA developer resources and GitHub. Scientists can access BioNeMo-powered workflows through Claude Science, which is entering public beta today. Anthropic is inviting researchers to provide feedback on additional domain specialists and integrations they need as part of the public beta.