An Altos Labs team using a flow-matching model wins the Generalist Prize in Arc's first Virtual Cell Challenge

Arc Institute/Virtual Cell Challenge

Arc Institute's first Virtual Cell Challenge, designed to predict how cells respond when genes are switched off, drew over 5,000 registrants from 114 countries and 1,200 teams submitting results. Altos Labs won the Generalist Prize with a flow-matching approach—a type of generative AI model—and published a preprint describing the method.

• Participation: 5,000+ registrants, 114 countries, 1,200 submitting teams, 300+ final submissions.
• Winner: Altos Labs team using flow-matching generative model.
• Prize: Generalist Prize for best cross-cellular generalization.
• Goal: advance in silico cellular prediction to rival breakthroughs in other AI domains.