Research

Benchmarks, papers, and writing from our team on improving AI agents for real-world healthcare environments.

Writing

Apr 14, 2026

Introducing HealthAdminBench: AI Agents Can Diagnose Rare Diseases, But Can They Handle Your Insurance?

Published at an ICLR 2026 Workshop, HealthAdminBench is the first benchmark for realistically evaluating LLM agents on the healthcare admin workflows that cost our economy $1 trillion/year.

Apr 13, 2026

Towards the Autonomous Hospital

Our mission is to advance frontier model capabilities for solving real-world, economically meaningful healthcare workflows.

Papers

May 2026Publication

MEDS: An emerging data standard and ecosystem for health AI research

NEJM AI 2026↗

Apr 2026Benchmark

HealthAdminBench: Can AI agents handle your insurance?

ICLR 2026 Workshop→

May 2025Benchmark

MedHELM: Holistic evaluation of large language models for medical tasks

Nature Medicine 2026↗

Apr 2025Publication

Context Clues: Evaluating long-context models on EHR data

ICLR 2025↗

Mar 2025Benchmark

Top of the CLASS: Benchmarking LLM agents on real-world enterprise tasks

ICLR 2025 Workshop↗

Jan 2025Publication

Zero-shot clinical trial patient matching with large language models

NEJM AI 2025↗

May 2024Publication

Automating the enterprise with foundation models

VLDB 2024↗

Feb 2024Benchmark

MedAlign: A clinician-generated dataset for instruction following with EHRs

AAAI 2024↗

Dec 2023Benchmark

EHRSHOT: An EHR benchmark for few-shot evaluation of foundation models

NeurIPS 2023↗

Nov 2023Benchmark

INSPECT: A multimodal dataset for patient outcome prediction

NeurIPS 2023↗

Oct 2023Publication

The shaky foundations of large language models for electronic health records

npj Digital Medicine 2023↗