Anthropic integrates NVIDIA BioNeMo into Claude Science to accelerate life sciences research
Anthropic has integrated the NVIDIA BioNeMo Agent Toolkit into its new Claude Science workbench to streamline computational life sciences research and accelerate data-intensive workflows.
1. Integration of BioNeMo and Claude Science
Anthropic has launched the public beta of Claude Science, an AI workbench designed to assist with computational life sciences research. The platform allows scientists to use natural language to interact with digital agents, which then execute complex research workflows. This system is integrated with the NVIDIA BioNeMo Agent Toolkit, enabling the platform to access high-performance computing resources, scientific models, and NVIDIA NIM microservices as callable skills. By translating natural language requests into operational actions, the system aims to streamline tasks such as genomic analysis, protein structure prediction, and molecular design.
2. Enhancing Research Efficiency
The integration provides researchers with access to advanced biomolecular models, including Evo 2, Boltz-2, and OpenFold3. By utilizing NVIDIA’s accelerated software libraries, the system significantly reduces the time required for data-intensive processes. For example, the RAPIDS-singlecell tool can reduce preprocessing and clustering workflows from 52 minutes to 25 seconds, while the nvMolKit accelerates cheminformatics tasks by up to 3,000 times. These capabilities allow scientists to maintain an iterative loop between human reasoning and machine-accelerated processing, focusing on core research rather than manual software configuration.
3. Deployment and Industry Adoption
NVIDIA BioNeMo is already utilized by 18 of the top 20 global pharmaceutical companies. The platform uses containerized NIM microservices to provide stable, enterprise-ready inference endpoints for production environments. The BioNeMo Agent Toolkit is designed to be harness-agnostic, ensuring that scientific skills remain consistent across different agent frameworks and research platforms. During the current public beta phase, Anthropic is soliciting feedback from the scientific community to refine the platform’s integrations and expand its domain-specific capabilities.