Introducing
An applied research group advancing healthcare AI — responsibly, transparently, and with the rigor it demands.
Proud partner of
Healthcare AI needs intelligence that’s built for healthcare.
Most AI is built for general-purpose use. It wasn’t trained on clinical populations, evaluated on healthcare outcomes, or designed for the privacy constraints that healthcare data demands. The result is models that work well on general benchmarks but break down in practice.
The tools to build member and patient experiences safely, models that ensure accuracy, and standards that let you know if things are actually working — this our mandate.
This is what building for healthcare actually looks like.
Privacy-preserving data
Training AI on real patient data risks exposing the most sensitive information people have. We built SYNTH — a platform that generates synthetic patients that mirror the diversity and clinical complexity of real populations, without exposing a single medical record.
Purpose-built models
General-purpose AI wasn’t designed for healthcare. We’re building small language models purpose-built for clinical use — private, fast, affordable, and more accurate where it counts. more private, faster, more affordable, and more accurate where it counts.
Patient-centered benchmarks
There’s no shared standard for what “great” looks like in healthcare AI. We’re building the benchmarks that change that — evaluating AI performance across the full health journey, from understanding options to finding the right care at the right time.
Our first dataset is free.
We’re publishing our first SYNTH-generated dataset on Hugging Face, open to the entire research community. It reflects the real-world distribution of health conditions across League’s platform, giving researchers the diversity and depth they need to build better healthcare AI.
A decade of healthcare intelligence. Now a foundation for everyone.
League has spent over a decade building AI infrastructure at the intersection of payers, providers, and patients — accumulating billions of real interactions across 70 million people around the world. That depth, spanning care navigation, clinical signal processing, and personalized health engagement at scale, is the foundation League Labs is built on.
And we’re turning it over to the field. Papers, benchmarks, open-source releases — because the future of healthcare AI shouldn’t be locked inside any one company or country.