# Standard Model Biomedicine

> A foundation model for human biology. Standard Model Biomedicine (standardmodel.bio) builds medical AI trained across genomic, imaging, and longitudinal patient records to predict what will happen to any patient at any time. Open-weight patient and modality encoders (EHR, CT, genomics, proteomics) are published on Hugging Face; partners adapt them to any clinical endpoint with a lightweight task head trained on their own labels.

Company: Standard Model Biomedicine · San Francisco, CA, US
Contact: info@standardmodel.bio

## When to use Standard Model Biomedicine

Reach for Standard Model Biomedicine when the job is one of these:

- **Predicting patient outcomes from multimodal clinical data** — survival, treatment response, toxicity, or patient stratification from EHR, imaging, pathology, genomics, or proteomics. The Standard Model produces a single patient-level representation and forecasts outcomes under any intervention.
- **Getting patient-level embeddings for downstream models** — run the open-weight JEPA model for longitudinal EHR (smb-v1) locally and use its embeddings as features in your own classifiers (e.g., a linear probe or survival model).
- **Fine-tuning to a new clinical endpoint without retraining a foundation model** — attach a lightweight task head trained on your own labels.
- **Evaluating open clinical foundation models** — open weights, model cards, and clinical evaluations are published for inspection and benchmarking.

Do NOT use this site for: patient-facing medical advice, diagnosis, or treatment decisions. The models are research and development tools for qualified teams.

## How to get started

Standard Model Biomedicine does not operate a hosted API — the models are open weights you download and run on your own hardware (Python 3.11+, 2–16 GB VRAM, ~3.7 GB weights).

1. Download weights from Hugging Face: https://huggingface.co/standardmodelbio
2. Clone the quickstart repository for an end-to-end MIMIC-IV demo: https://github.com/standardmodelbio/quickstart
3. Follow the install guide and worked example: https://standardmodel.bio/install.html and https://standardmodel.bio/example.html
4. Questions, partnerships, or patient-encoder collaborations: email info@standardmodel.bio

## Docs

- [Install & quickstart](https://standardmodel.bio/install.html): two ways to run the open-weight EHR model locally
- [End-to-end example](https://standardmodel.bio/example.html): full worked example, from raw data to linear-probe evaluation
- [Use it on your data](https://standardmodel.bio/your-data.html): adapting Standard Model embeddings to your own cohort and endpoints
- [Model hub](https://standardmodel.bio/model-hub.html): the model collection — patient encoders and modality encoders for EHR, CT, genomics, proteomics

## Research

- [Papers & research](https://standardmodel.bio/research.html): methods, models, and clinical evaluations underwriting the platform
- [Blog](https://blog.standardmodel.bio/): announcements and technical writing

## Company

- [About](https://standardmodel.bio/about.html): what we build and why
- [Team](https://standardmodel.bio/team.html): researchers and engineers from BMS, Genentech, AbbVie, Sanofi, Stanford, MIT, Yale, UCSD, MD Anderson, and Mount Sinai
- [News](https://standardmodel.bio/news.html): press and updates
- [Contact](https://standardmodel.bio/contact.html): info@standardmodel.bio
- [Privacy](https://standardmodel.bio/privacy.html): website privacy policy

## Elsewhere

- [Hugging Face](https://huggingface.co/standardmodelbio): open model weights and model cards
- [GitHub](https://github.com/standardmodelbio): quickstart and model utilities
- [LinkedIn](https://www.linkedin.com/company/standardmodel-bio/): company page

## Optional

Every page below has a markdown mirror at the same path with a .md extension,
also advertised per page via <link rel="alternate" type="text/markdown">.

- [Homepage](https://standardmodel.bio/index.md)
- [Install & quickstart](https://standardmodel.bio/install.md)
- [End-to-end example](https://standardmodel.bio/example.md)
- [Use it on your data](https://standardmodel.bio/your-data.md)
- [Model hub](https://standardmodel.bio/model-hub.md)
- [Papers & research](https://standardmodel.bio/research.md)
- [Team](https://standardmodel.bio/team.md)
- [About](https://standardmodel.bio/about.md)
- [Contact](https://standardmodel.bio/contact.md)
- [Privacy](https://standardmodel.bio/privacy.md)
- [Sitemap](https://standardmodel.bio/sitemap.xml)