benchmarking

AlphaFold made proteins predictable; the field’s next moonshot is a virtual cell you can query — give it a starting cell state and a perturbation, and have it predict the gene-expression response you’d otherwise run an experiment to see. Arc Institute has turned that dream into a CASP-style contest, the Virtual Cell Challenge, drawing 5,000+ entrants, and shipped its first-generation State model. But the same year delivered a bracing reality check: a Nature Methods benchmark found that today’s single-cell foundation models — scGPT, Geneformer and peers — still don’t beat simple linear baselines at predicting perturbation effects. It’s our flagship bet, seen honestly: a thrilling ambition and the scoreboard that will keep it accountable.

Genomic foundation models had their victory year; 2026 is the audit. Independent held-out benchmarks now separate what these models can demonstrate from what survives an honest baseline — and the useful answer isn’t one model to rule them all, but the right model for each variant class, kept open and interpretable.