Biomni: A General-Purpose Biomedical AI Agent for Research

Summary

Biomni is a general-purpose biomedical AI agent designed to autonomously execute a wide range of research tasks across diverse biomedical subfields. It integrates cutting-edge large language model reasoning with retrieval-augmented planning and code-based execution. This powerful tool helps scientists significantly enhance research productivity and generate testable hypotheses.

Use at your own risk

OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.

Introduction

Biomni is an innovative, general-purpose biomedical AI agent developed by SNAP Stanford. It is designed to autonomously execute a wide range of research tasks across diverse biomedical subfields, significantly enhancing research productivity. By integrating cutting-edge large language model (LLM) reasoning with retrieval-augmented planning and code-based execution, Biomni empowers scientists to generate testable hypotheses and streamline complex workflows.

Installation

To get started with Biomni, follow these steps for installation and environment setup.

First, set up the environment by following the instructions in the biomni_env/README.md file within the repository.

Then, activate the environment:

conda activate biomni_e1

Install the Biomni pip package:

pip install biomni --upgrade

For the latest updates, install directly from the GitHub source:

pip install git+https://github.com/snap-stanford/Biomni.git@main

Finally, configure your API keys for LLM providers like Anthropic or OpenAI. You can do this by creating a .env file or setting shell environment variables, as detailed in the official Biomni repository.

Examples

Once installed and configured, Biomni can be used to execute biomedical tasks using natural language.

Here are some basic usage examples:

from biomni.agent import A1

# Initialize the agent with data path, Data lake will be automatically downloaded on first run (~11GB)
agent = A1(path='./data', llm='claude-sonnet-4-20250514')

# Execute biomedical tasks using natural language
agent.go("Plan a CRISPR screen to identify genes that regulate T cell exhaustion, generate 32 genes that maximize the perturbation effect.")
agent.go("Perform scRNA-seq annotation at [PATH] and generate meaningful hypothesis")
agent.go("Predict ADMET properties for this compound: CC(C)CC1=CC=C(C=C1)C(C)C(=O)O")

Biomni also provides a Gradio-based web interface for interactive use. You can launch it with:

from biomni.agent import A1

agent = A1(path='./data', llm='claude-sonnet-4-20250514')
agent.launch_gradio_demo()

Remember to install Gradio separately: pip install "gradio>=5.0,<6.0".

Why Use Biomni?

Biomni offers several compelling reasons for researchers and developers in the biomedical field:

General-Purpose AI Agent: It can handle a wide array of research tasks across various biomedical subfields, from experimental planning to data analysis.
LLM-Powered Reasoning: Leverages advanced large language models for intelligent planning, retrieval-augmented generation, and code-based execution.
Enhanced Productivity: Automates complex, multi-step research workflows, allowing scientists to focus on higher-level problem-solving.
Know-How Library: Integrates a curated collection of best practices, protocols, and troubleshooting guides, providing domain expertise on demand.
Biomni-R0 Model: Features a specialized reasoning model, Biomni-R0, built on Qwen-32B, optimized for tool use and complex biological problem-solving.
Biomni-Eval1 Benchmark: Provides a comprehensive evaluation benchmark for assessing biological reasoning capabilities across 10 diverse tasks.
Community-Driven Development: Welcomes contributions for new tools, datasets, software integrations, benchmarks, and know-how documents, fostering an open-science ecosystem.