Biological models
Understand model architectures, biological inputs, access requirements and the evidence from linked evaluations.
318 model records in release 2026-10-23-5347695500ed (release manifest, gzip). Showing 313–318; page 14 of 14.
Evaluated configurations, methods and pipelines are listed separately. Names alone do not establish equivalent models or checkpoints.
VESPAl
Proteins and complexes
Light version of VESPA that predicts single amino acid variant effects from predicted conservation and BLOSUM62 scores only, without the language-model substitution probabilities, scored in ProteinGym as a zero-shot substitution baseline.
ViroCaduceus
Microbes and communities · DNA and genomes
Caduceus-architecture model pretrained on viral sequences by the ViroBench authors (in-house model).
ViroDNABERT2
Microbes and communities · DNA and genomes
DNABERT-2-architecture model pretrained on viral sequences by the ViroBench authors (in-house model).
ViroHyena
Microbes and communities · DNA and genomes
Hyena-architecture nucleotide model pretrained on viral sequences by the ViroBench authors (in-house model).
Wavenet
Proteins and complexes
Alignment-free autoregressive deep generative model of protein sequences adapted from natural language processing that predicts missense and indel effects, scored in ProteinGym as a zero-shot substitution baseline.
xTrimoPGLM
Proteins and complexes
Family of protein language models, scaled up to 100 billion parameters, pretrained jointly with autoencoding and autoregressive objectives for protein understanding and generation; ProteinGym scores several sizes as zero-shot single-sequence substitution baselines.
This index reflects a dated catalogue, not an exhaustive census. Source checking does not mean independent reproduction; compare results only under compatible protocols, datasets and metrics.