All Competitors

A 4.7B-parameter Mixture-of-Experts genomic foundation model pretrained on ~1.2 trillion nucleotide tokens from human-associated microbial genomes.

A permutation-invariant denoising transformer trained on ~2 million bacterial community samples to learn member-level stability scores and predict microbiome composition dynamics zero-shot.

A family of GPT-2-style causal language models pretrained on 539,000+ microbiome samples, enabling zero- and few-shot transfer across microbiome prediction tasks.