The Arc Virtual Cell Challenge invites participants to develop models that predict the impact of gene silencing on cell behavior using CRISPR technology. With a curated dataset of approximately 300,000 single-cell RNA profiles, the challenge emphasizes the importance of context generalization in machine learning for biological applications while providing foundational biological knowledge for participants from other fields.