Biomni: Stanford's open-source biomedical AI agent goes live for 10,000 scientists
Synopsis
Stanford’s Biomni, published in Science, is the first general-purpose biomedical AI agent that autonomously runs entire research workflows from a plain-text prompt — and 10,000 scientists are already using it for free.
Key Takeaways
Biomni is described as the first general-purpose biomedical AI agent, developed by a Stanford University -led team including researchers Yuanhao Qu and Kexin Huang .
The system has been released as an open-source platform with a web interface, requiring no coding skills from end users.
More than 10,000 scientists worldwide are already using Biomni for everyday research tasks, according to supervising professor Jure Leskovec .
The research was published in the peer-reviewed journal Science in the week of July 10, 2026 .
Biomni successfully generated lab protocols that scientists followed in real experiments, spanning domains including cancer biology , DNA analysis, and molecular cloning .
In one benchmark test, the agent autonomously processed wearable-device data files, cleaned them, ran analysis, and produced new biological hypotheses without human intervention.
Stanford University researchers have unveiled Biomni, described as the first general-purpose biomedical AI agent capable of collaborating with human scientists on complex research tasks — and it is already in active use by more than 10,000 scientists worldwide. The system, detailed in the journal Science this week, converts plain-language requests into full research workflows, handling everything from database searches to laboratory protocol generation.
What Biomni does
According to the research team, Biomni can take a simple text prompt and autonomously execute an end-to-end scientific pipeline — searching biomedical databases, writing analysis code, identifying disease-causing genes, and producing step-by-step lab instructions that scientists have successfully followed in real experiments. The system is built as an open-source platform with a web interface, meaning biologists can deploy it without any programming knowledge.
In one documented test, Biomni was handed hundreds of raw data files from wearable devices and tasked with finding biological patterns. It independently cleaned the data, ran statistical analyses, and generated novel hypotheses — a workflow that would traditionally require a multidisciplinary team of specialists.
The team behind it
Jure Leskovec, a Stanford computer science professor who supervised the project, confirmed the release on Tuesday, July 8, 2026. The team includes researchers Yuanhao Qu and Kexin Huang, both with roots in Beijing, alongside collaborators across San Francisco-area institutions. “We have over 10,000 scientists all over the world using the system for their everyday tasks,” Leskovec said.
Why it matters
Biomedical research has historically been bottlenecked by the need for cross-disciplinary expertise — a molecular biologist may lack the computational skills to analyse genomic data at scale, while a data scientist may lack the domain knowledge to interpret results. Biomni is positioned to collapse that gap, automating tasks that span cancer biology, DNA analysis, molecular cloning, and wearable-device data interpretation under a single agent.
The publication in Science — one of the most rigorously peer-reviewed journals in the world — lends the claims significant institutional credibility, distinguishing Biomni from the wave of AI research tools that have circulated without formal validation.
The competitive backdrop
The release arrives amid intensifying competition in AI-for-science, with major technology companies and well-funded startups racing to automate laboratory workflows. Tools such as Phylo have addressed narrower genomics tasks, but a truly general-purpose agent spanning the full biomedical stack has remained elusive. Biomni's open-source model also sets it apart from proprietary platforms, potentially accelerating adoption in academic and public-health settings globally.
What’s next
With 10,000 active users already on the platform, the research team’s next challenge will be scaling infrastructure, ensuring reproducibility across diverse experimental contexts, and navigating regulatory questions around AI-generated lab protocols — particularly in sensitive domains such as research involving human embryos. The trajectory of adoption over the next two quarters will be a key signal of whether general-purpose biomedical agents can move from academic novelty to mainstream research infrastructure.
Point of View
Backed by peer review rather than marketing claims. What mainstream coverage underplays is the geopolitical subtext — a Stanford-led team with prominent Chinese researchers producing foundational open-source biomedical infrastructure at a moment when US-China scientific collaboration is under sustained political pressure. The open-source release is also a strategic move: by giving away the platform, the team builds a global user base that generates the feedback loops and real-world validation data that proprietary competitors cannot easily replicate. The harder question — whether AI-generated lab protocols can meet regulatory and reproducibility standards in clinical or drug-discovery pipelines — remains unanswered and will define the technology’s ceiling.
NationPress
10 Jul 2026
Frequently Asked Questions
What is Biomni and what can it do?
Biomni is an open-source, general-purpose biomedical AI agent developed by a Stanford University -led team and published in the journal Science in July 2026. It converts plain-language text prompts into complete research workflows, autonomously searching databases, writing code, identifying disease genes, and generating laboratory protocols that scientists have executed in real experiments.
Is Biomni free to use?
Yes. Biomni has been released as a free, open-source system with a web interface, meaning researchers can access it without writing any code. According to professor Jure Leskovec , more than 10,000 scientists worldwide were already using it as of Tuesday, July 8, 2026 .
Why does Biomni matter for biomedical research?
Biomni matters because it automates tasks that previously required teams of specialists with overlapping expertise in data science, genomics, and laboratory biology. Its peer-reviewed validation in Science — covering areas from cancer biology to molecular cloning — distinguishes it from the many unverified AI research tools currently on the market.
How does Biomni compare to other AI science tools?
Unlike narrow tools such as Phylo , which address specific genomics tasks, Biomni is designed as a general-purpose agent spanning the full biomedical research stack. Its open-source model also contrasts with proprietary AI platforms, potentially giving it a faster adoption curve in academic and public-health institutions globally.
