Predicting cell state in previously unseen conditions such as disease or in response to a drug has typically required retraining for each new biological context. Today, Arc is releasing Stack, a foundation model that learns to simulate cell state under novel conditions directly at inference time, no fine-tuning required.

Stack captures something that most models miss: cellular context. A T cell in inflamed tissue behaves differently, not just because of its own genes, but because of its environment. Stack processes cells together & learns from those relationships.

Just as text prompts guide language models, cells serve as prompts in Stack. It can observe drug-treated immune cells & predict how epithelial cells would respond to the same drug, a task never explicitly trained for. It's the first single-cell foundation model capable of in-context learning, or generalizing to new tasks during inference.

The team applied Stack to build Perturb Sapiens: An atlas of ~20,000 predicted cell responses across 28 tissues & 201 perturbations + a subset validated using held-out datasets, confirming that predictions capture real biology.