ESM-2 / ESMFold vs ProtTrans

ESM-2 / ESMFold: Protein language models and single-sequence structure prediction (Meta repo archived; models still widely used via Hugging Face). ProtTrans: Transformer protein language models (ProtT5, ProtBERT and others) pretrained on UniRef and BFD, used for embeddings and fine-tuning.

Facts below come from each project's own website, repository and paper. We don't publish benchmark claims here; see the papers for evaluations.

FieldESM-2 / ESMFoldProtTrans
TaskProtein language model, structure predictionProtein embeddings / property prediction
ModalitiesProtein (single chain)Sequence
MSA requiredNo (single sequence)No
Open weightsYesYes (MIT)
GPU need—GPU recommended; CPU works for small models
LicenseMITMIT
PricingOpen sourceOpen source
Released2022-07-212020-05-11
PaperLinkLink
GitHubGitHubGitHub
GitHub stars4,1701,324
InterfacePackage, Colab—

License and commercial use

FieldESM-2 / ESMFoldProtTrans
Code licenseMITMIT
Weights license—MIT
Commercial usePermissive (commercial OK)Permissive (commercial OK)
AnswerYes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices. DetailsYes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices. Details

Add more Protein structure & design tools to this comparison →

More Protein structure & design comparisons

← All Protein structure & design tools