Genomics & Single-Cell · Arc Institute / Stanford
Genome-scale foundation model spanning nucleotide to whole-genome context.
5–30min
Runs on Managed GPU endpoint
Runs on managed compute. Availability and startup time depend on current capacity.Run this model
About this model
Evo 2 is trained across the tree of life on multi-kingdom DNA with a context window long enough to reason about whole genomic loci. It both generates and scores sequences, which makes it usable for variant effect prediction and for designing regulatory elements.
Standardized I/O contract
Every model in the Hub speaks the same contract, which is what lets the Router and the agent call any of them without special-casing.
Inputs
DNA sequence
Tokens to generate
Outputs
Continuation sampled from the model.
Model likelihood, usable as a variant effect score.
