All Competitors

A Dirichlet flow-matching model that generates family-aware protein sequences by initializing from ancestral-reconstruction lineage priors rather than random noise.

Leakage-controlled multimodal model predicting within-species relative protein expression across 385 bacterial species, with transfer to unseen phyla.

OpenAI's first life-sciences frontier reasoning model, optimized for multi-step scientific workflows spanning protein engineering, genomics, drug-target discovery, and biochemistry reasoning.

Edit-based flow-matching model that generates variable-length protein variants via learned insertions, deletions, and substitutions on a template sequence.

Transformer for few-shot protein fitness prediction that combines in-context learning with test-time training to adapt to new proteins and assays.

Codon-level BERT model that captures genomic signals invisible to amino acid models, outperforming billion-parameter PLMs with just 86M parameters.