Atlas 2

Gold definitionUpdated Apr 2, 2026

Definition

Atlas 2 is a pathology vision foundation model designed to overcome limitations of previous models, achieving state-of-the-art performance in prediction, robustness, and resource efficiency. It was trained on the largest pathology dataset to date, comprising 5.5 million histopathology whole slide images.

At a glance

Executive summary

Atlas 2 is a new AI model for analyzing pathology images, designed to be highly accurate, reliable, and efficient. It was trained on a massive dataset of 5.5 million medical images, allowing it to outperform previous models and be more suitable for use in hospitals and research.

TL;DR

Atlas 2 is a cutting-edge AI model for pathology that uses a huge dataset to achieve top performance, robustness, and efficiency for clinical use.

Key points

Trained on the largest pathology foundation model dataset to date (5.5 million histopathology WSIs).
Solves the problem of performance, robustness, and computational tradeoffs in previous pathology foundation models.
Used by computational pathology researchers and potentially in clinical diagnostic settings at institutions like Charité, LMU Munich, and Mayo Clinic.
Unlike prior pathology foundation models, Atlas 2 achieves state-of-the-art performance across prediction, robustness, and resource efficiency.
Represents a trend towards highly specialized and robust foundation models for specific medical imaging domains.

Use cases

Automated cancer diagnosis and grading from histopathology slides.
Accelerating drug discovery by analyzing tissue responses in preclinical studies.
Personalized medicine through precise biomarker identification from biopsy images.
Research into disease mechanisms by enabling large-scale quantitative analysis of tissue morphology.
Improving efficiency in pathology labs by assisting pathologists with initial screening and quantification tasks.

Also known as

Atlas 2-B, Atlas 2-S