NVIDIA has announced the integration of its BioNeMo Agent Toolkit with Anthropic's newly launched Claude Science, enabling researchers to use natural language to perform complex scientific workflows across genomics, drug discovery, protein engineering and molecular design.
The collaboration marks a significant step towards agentic AI for scientific research, allowing scientists to move from research intent to computational execution without manually configuring models, software environments or computing infrastructure. By integrating NVIDIA's GPU-accelerated models, libraries and NVIDIA NIM microservices into Claude Science, researchers can seamlessly access high-performance AI tools within a unified research environment.
Scientists can simply describe research objectives—such as analysing genomic sequences, predicting protein structures or designing potential drug binders—in natural language. Claude Science then orchestrates specialised AI agents that leverage NVIDIA BioNeMo capabilities, including models such as Evo 2, Boltz-2 and OpenFold3, to execute accelerated workflows across genomics, proteomics, single-cell analysis, cheminformatics and clinical research.
According to NVIDIA, the platform creates an iterative research cycle where scientists can review computational results, refine hypotheses and initiate subsequent analyses while remaining focused on scientific discovery. One highlighted application involves designing inhibitors for cancer-related targets, where AI agents rapidly predict, optimise and validate potential drug candidates, significantly accelerating early-stage drug discovery.
The announcement reinforces NVIDIA's growing presence in the life sciences sector. The company stated that 18 of the world's top 20 pharmaceutical companies now use NVIDIA BioNeMo technologies to support AI-driven research spanning genomics, medical imaging, protein engineering and molecular modelling.
The BioNeMo Agent Toolkit also incorporates high-performance tools such as NVIDIA Parabricks for rapid genomic analysis, RAPIDS-singlecell for accelerated single-cell workflows, and nvMolKit, which speeds up cheminformatics operations by up to 3,000 times. Designed as an open, framework-agnostic platform, the toolkit is available through NVIDIA's developer resources and GitHub, enabling researchers and developers to build interoperable AI agents for next-generation scientific discovery.
