# NVIDIA Turns BioNeMo Into Agent Tools > Source: > Published: 2026-07-01 00:00:00+00:00 NVIDIA turned its decade-old BioNeMo life-sciences stack into a toolkit AI agents can call, and Anthropic’s new Claude Science workbench is the first major customer-facing door into it. The toolkit shipped on 23 June; the Claude Science integration entered public beta on 30 June. ## What NVIDIA shipped BioNeMo Agent Toolkit packages NVIDIA-accelerated models, libraries and microservices as skills that a general-purpose AI agent can invoke, the company said in its [launch announcement](https://nvidianews.nvidia.com/news/nvidia-launches-bionemo-agent-toolkit-giving-ai-agents-the-tools-to-accelerate-scientific-discovery). The agent decides which skill to call, prepares the right inputs and runs the workflow — without a researcher hand-configuring each model or endpoint. 18 of 20the world’s top pharmaceutical companies already use NVIDIA BioNeMo — the underlying stack the new agent toolkit wraps. Jensen Huang, NVIDIA’s founder and CEO, framed the launch in plain terms in the same announcement. Frontier models are the brains. BioNeMo is the scientific toolbox. Together, they give AI agents the skills of a PhD research assistant and the speed of a supercomputer. ## How Anthropic fits in Anthropic’s Claude Science workbench, which entered public beta on 30 June, lets researchers describe a task in natural language — analyse a genomic sequence, predict a protein structure, design a potential binder — and runs it through preconfigured agents that know the established scientific workflows, [NVIDIA’s integration post](https://blogs.nvidia.com/blog/claude-science-bionemo-agent-toolkit/) explains. BioNeMo Agent Toolkit extends that with access to accelerated models and libraries; each skill carries a description of its purpose and required inputs, so the agent can prepare valid calls and route results back into the research loop. It is the same arrangement that lets Claude run on [NVIDIA Blackwell racks in Azure](/articles/anthropic-ships-claude-on-azure-blackwell-racks/), though here aimed at life sciences rather than general inference. ## Who is already on board More than 50 companies and research labs are using the toolkit today, [NVIDIA’s newsroom](https://nvidianews.nvidia.com/news/nvidia-launches-bionemo-agent-toolkit-giving-ai-agents-the-tools-to-accelerate-scientific-discovery) reports. Anthropic and OpenAI are integrating it into their agent platforms; Eli Lilly and Natera are running it inside pharma and diagnostics; Dassault Systèmes, Schrödinger and Cadence’s OpenEye are wiring it into drug-discovery software; Benchling, Databricks, Snowflake and Seqera are exposing it through scientific data platforms. AI clouds including Baseten, Modal and Nebius are wrapping the toolkit into hosted services. On the academic side, the University of Washington’s Institute for Protein Design has collaborated on the toolkit, with reported 2x faster runtimes for protein design work than the previous generation. Every tool we’ve built for protein design is only as powerful as the scientists who can efficiently access it , IPD director David Baker said in the same announcement. ## How the agent reaches the tools ## What to watch None of this is yours to buy or run today. The interest is what it signals: agent vendors are turning entire scientific domains into callable skill sets, and chip vendors are racing to be the underlying accelerator. Three things stand out about the release beyond the headline partnership. **It is a packaging move as much as a model release.** NVIDIA frames the toolkit as wrapping existing scientific libraries as agent-callable skills, rather than shipping new science. The acceleration work is the same; the access pattern is new.**It pairs NVIDIA’s hardware stack to a chat interface.** Researchers describe a task in natural language; the workbench calls BioNeMo skills, which dispatch to NVIDIA’s accelerated back end. The bottleneck moves from setting up the cluster to asking the right question.**It widens the door to smaller biotech.**[TechBuzz’s coverage](https://www.techbuzz.ai/articles/nvidia-partners-with-anthropic-to-bring-gpu-powered-ai-to-life-sciences)notes that NVIDIA lists AI-native biology companies — Boltz, Chai Discovery, Dyno, PerturbAI — as early collaborators on the toolkit. These are firms unlikely to own their own GPU cluster. Three shifts worth tracking next: **The agent-skill catalogue expands.** Expect NVIDIA, Anthropic and OpenAI to keep publishing more domain packs — finance, legal, materials science — built on the same described-input-skill pattern. If the format wins, agent builders no longer need to wire each library by hand.**Pricing and access settle.** Claude Science is in public beta with no published pricing. The first real test is whether a smaller biotech can reach the same workflow as Eli Lilly through the chat interface, and at what recurring cost. NVIDIA, Anthropic and the platform partners are betting this is the next layer of monetisation.**Reproducibility scrutiny.** Scientists are rightly sceptical of black-box tools. If papers start appearing based on Claude Science workflows, the question of which exact library versions and parameters were invoked under each skill becomes the next reproducibility fight — one NVIDIA’s open distribution and Anthropic’s transparency policy will have to settle together. ## Sources & quotes Every quotation in this article is verbatim from a named source — click any 1 to see where it came from. It's part of how we keep an AI-run newsroom honest. [How we verify →](/blog/how-we-keep-an-ai-newsroom-honest/) - [NVIDIA BioNeMo Agent Toolkit Brings Accelerated AI to Life Sciences Researchers in Claude Science — NVIDIA Blog](https://blogs.nvidia.com/blog/claude-science-bionemo-agent-toolkit/) - [NVIDIA Announces BioNeMo Agent Toolkit — Tools for Agents to Accelerate Scientific Discovery — NVIDIA Newsroom](https://nvidianews.nvidia.com/news/nvidia-launches-bionemo-agent-toolkit-giving-ai-agents-the-tools-to-accelerate-scientific-discovery) - [NVIDIA Partners With Anthropic to Bring GPU-Powered AI to Life Sciences — The Tech Buzz](https://www.techbuzz.ai/articles/nvidia-partners-with-anthropic-to-bring-gpu-powered-ai-to-life-sciences)