Anthropic has launched the public beta of Claude Science, a new AI research platform designed to help scientists automate complex computational workflows using natural language. A key part of the platform is its native integration with NVIDIA’s BioNeMo Agent Toolkit, giving researchers direct access to powerful GPU-accelerated computing resources without requiring specialized technical expertise.
The collaboration aims to simplify scientific research by allowing researchers to describe their goals in plain language while AI agents handle the computational work behind the scenes.
Natural language meets scientific computing
Claude Science functions as an AI-powered research assistant capable of translating written instructions into fully automated scientific workflows.
Instead of manually configuring machine learning models, setting up software environments, or managing computational infrastructure, scientists can simply ask Claude Science to perform tasks such as analyzing genomic sequences, predicting protein structures, or designing new molecular compounds.
The platform interprets these requests and automatically coordinates the necessary NVIDIA-powered tools to complete the analysis.
NVIDIA’s BioNeMo ecosystem
The integration is built on NVIDIA’s extensive BioNeMo ecosystem, which combines GPU hardware, AI frameworks, scientific models, optimized software libraries, and enterprise-ready microservices.
These resources provide researchers with accelerated computing for a wide variety of life sciences applications while dramatically reducing processing times.
According to NVIDIA, 18 of the world’s 20 largest pharmaceutical companies already use BioNeMo within production environments, highlighting its growing importance across the pharmaceutical industry.
AI agents automate research workflows
The BioNeMo Agent Toolkit gives Claude Science access to specialized AI agents trained for different scientific disciplines, including genomics, proteomics, cheminformatics, clinical research, and single-cell analysis.
Each agent understands established scientific workflows and can determine which computational tools are required for a particular task.
For example, if a researcher wants to design potential treatments for a cancer mutation, Claude Science can coordinate multiple AI models to identify biological targets, generate candidate molecules, optimize their structures, and evaluate their potential effectiveness before presenting the results for human review.
This creates a continuous feedback loop where scientists refine their ideas while AI rapidly performs the computational work.
Faster genomic and molecular analysis
One of the platform’s biggest advantages is the dramatic reduction in processing time for computational biology tasks.
The toolkit includes access to advanced AI models such as Evo 2, Boltz-2, and OpenFold3, alongside optimized NVIDIA software designed specifically for life sciences.
Genomic analysis performed through NVIDIA Parabricks can be reduced from several hours to just minutes, allowing researchers to incorporate genomic insights into experiments almost immediately.
Similarly, RAPIDS-singlecell, developed by scverse, reduces preprocessing and clustering of a dataset containing 1.3 million cells from approximately 52 minutes to only 25 seconds.
For cheminformatics, nvMolKit accelerates tasks including molecular similarity searches and conformer generation by as much as 3,000 times, enabling researchers to explore massive chemical libraries far more efficiently.
Enterprise-ready deployment
To support production-scale research environments, NVIDIA packages its biomolecular AI models as BioNeMo NIM microservices.
These fully containerized services provide optimized inference endpoints that organizations can deploy across enterprise infrastructure. AI agents communicate with these services through stable APIs, allowing complex scientific workflows to scale without requiring researchers to manage the underlying infrastructure.
The BioNeMo Agent Toolkit is also framework-agnostic, enabling organizations to integrate the same scientific capabilities into different AI agent systems and research platforms.
Accelerating the future of drug discovery
The launch of Claude Science reflects the growing convergence of generative AI and computational biology.
By combining Anthropic’s conversational AI with NVIDIA’s accelerated scientific computing platform, researchers can spend less time managing software and infrastructure while focusing more on scientific discovery.
As AI continues to play a larger role in life sciences, platforms like Claude Science may significantly shorten research timelines, improve experimental efficiency, and accelerate the development of new medicines and therapies.
