Biological models

318 model records in release 2026-10-23-5347695500ed (release manifest, gzip). Showing 241–264; page 11 of 14.

Evaluated configurations, methods and pipelines are listed separately. Names alone do not establish equivalent models or checkpoints.

• Protenix

  ByteDance biomolecular structure predictor built as a comprehensive open-source reproduction of AlphaFold 3, with released weights and training data pipeline.

• ProtGPT2

  Proteins and complexes

  Deep unsupervised language model trained on protein sequences that generates de novo protein sequences, scored in ProteinGym as a zero-shot single-sequence substitution baseline.

• Protriever

  Proteins and complexes

  End-to-end differentiable framework that learns to retrieve homologous protein sequences by vector search while training for fitness prediction, scored in ProteinGym as a zero-shot substitution baseline.

• ProtRNA

  RNA and transcriptomes

  ProtRNA RNA language model adapted from a protein language model.

• ProtSSN

  Proteins and complexes

  Pre-trained model that combines sequential and geometric encoders of protein primary and tertiary structure to score variant effects zero-shot, a ProteinGym substitution baseline from sequence and structure.

• ProtTrans

  Molecular interactions · Proteins and complexes

  ProtTrans protein language model family (T5-based).

• Puffin-D

  DNA and genomes

  Puffin-D transcription initiation model.

• Qwen 2.5 72B

  DNA and genomes

  72.7-billion-parameter causal language model in the Qwen2.5 series, pretrained on 18 trillion tokens, with a 131,072-token context.

• Qwen-2-7B

  Cells and tissues

  Language model in the Qwen2 series, a suite of foundational and instruction-tuned language models from 0.5 to 72 billion parameters.

• Qwen3.5-2B

  Cells and tissues

  Two-billion-parameter causal vision-language model in the Qwen3.5 series, with a hybrid Gated DeltaNet and gated attention layout and a native context of 262,144 tokens.

• Qwen3.6-27B

  Cells and tissues

  Dense 27-billion-parameter large language model in the Qwen3.6 series.

• RFamLlama

  RNA and transcriptomes

  RFamLlama autoregressive RNA language model trained on Rfam families (base and large).

• RFdiffusion

  Protein structure

  RFdiffusion generates protein structures, optionally conditioned on a motif, target or symmetry constraint.

• RhoFold+

  RNA and transcriptomes

  RhoFold+ predicts RNA three-dimensional structures from RNA sequence with alignment information.

• RhoMax

  Proteins and complexes

  Graph neural network that predicts the maximum absorption wavelength of a microbial rhodopsin from its AlphaFold2-predicted 3D structure.

• RiNALMo

  Microbes and communities · DNA and genomes

  RiNALMo RNA language model.

  Record: lit-microbial-20261011-virobench-model-rinalmo

• RiNALMo

  RNA and transcriptomes

  RiNALMo RNA language model (micro, mega and giga checkpoints).

  Record: lit-rna-20261011-rnalm-zeroshot-model-rinalmo

• RITA

  Proteins and complexes

  Suite of autoregressive generative transformer models for protein sequences, up to 1.2 billion parameters trained on UniRef-100, scored in ProteinGym as a zero-shot single-sequence substitution baseline.

• RNA-FM

  RNA and transcriptomes

  RNA-FM learns contextual representations of RNA nucleotides for downstream RNA analyses.

• RNA-km

  RNA and transcriptomes

  RNA-km RNA language model.

• RNA-MSM

  DNA and genomes

  RNA language model pre-trained on multiple sequence alignments.

• RNABERT

  DNA and genomes

  RNA language model pre-trained on small non-coding RNA.

• RNAErnie

  DNA and genomes

  RNA language model with motif-aware pre-training.

• RNAGenesis

  RNA and transcriptomes

  RNA model trained on non-coding RNAs that combines a BERT-style encoder with a diffusion-guided decoder to represent, predict structure for and design functional RNA sequences.

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.