ESMC - Biohub Platform

ESMC is the latest in the ESM family of protein language models. It advances representation learning for protein biology. Trained on billions of evolutionary sequences, it learns representations that reflect a mechanistic reduction of protein structure and function.

Hugging Face

GitHub

Paper

API reference

Get started

Get started using ESMC with our Quickstart and Tutorials.

Quickstart guide

Install the esm Python package

pip install esm

Create an API key

Generate an API key from your Biohub account. This API key manages your access to credits and tokens. The documentation often uses the terms API key and token interchangeably.

Connect to the Biohub Platform API

Call the ESM client with your selected model and replace <your API token> with your token name.

from esm.sdk.forge import ESMCForgeInferenceClient

client = ESMCForgeInferenceClient(model="esmc-6b-2024-12", url="https://biohub.ai", token="<your API token>")

Run your inference

Now you are ready to use your model. For examples of specific use cases or to append a sparse autoencoder (SAE) onto your ESMC model, check out our Tutorials.

Model tutorials

Embedding sequences with ESMC

Embed protein sequences and explore how different transformer layers encode structural and functional information.

Zero-shot entropy and mutation analysis

Compute per-position entropy and log-likelihood ratios to identify constrained vs. mutation-tolerant sites.

Layer sweep for enzyme function classification

Learn how to sweep all layers to find which one is best using enzyme classification as a task.

Understanding proteins with SAE features

Extract and visualize sparse autoencoder features, rank by peak activation, and map activations onto 3D structures.

Fine-tuning ESMC

Fine-tune a classification or regression head for your dataset on top of ESMC using parameter-efficient fine-tuning (PEFT).

Model details

For additional information, see the Hugging Face link.

Model card

Cite this model

Language Modeling Materializes a World Model of Protein Biology

@misc{candido2026language,
  title  = {Language Modeling Materializes a World Model of Protein Biology},
  author = {Candido, Salvatore and Hayes, Thomas and Derry, Alexander and Rao, Roshan
            and Lin, Zeming and Verkuil, Robert and Wu, Bryan and Lee, Jin Sub
            and Bruguera, Elise S. and Keval, Jehan A. and Kopylov, Mykhailo
            and Pak, John E. and Wu, Wesley and Thomas, Neil and Mataraso, Samson
            and Hsu, Alvin and Trotman-Grant, Ashton C. and Fatras, Kilian
            and dos Santos Costa, Allan and Badkundri, Rohil and Ak{\i}n, Halil
            and Oktay, Deniz and Deaton, Jonathan and Montabana, Elizabeth
            and Sitwala, Hrishita and Yu, Yue and Wiggert, Marius
            and Carlin, Dylan Alexander and Goering, Anthony W. and Blazejewski, Tomasz
            and Sandora, McCullen and Hla, Michael and Jia, Tina Z.
            and Kloker, Leon H. and Sofroniew, Nicholas J. and Uehara, Masatoshi
            and Pannu, Jassi and Bachas, Sharrol and Liu, Daniel S.
            and Sercu, Tom and Rives, Alexander},
  year   = {2026},
  url    = {https://www.biorxiv.org/content/10.64898/2026.06.03.729735},
  note   = {Preprint}
}

Overview
Details
Variants
Usage


Version	2026-04
Architecture	Transformer
Supported Modalities	Sequence
Training Data	Up to 6 billion proteins

Intended use

ESMC is designed for protein science research including structure prediction, function annotation, protein design, and understanding evolutionary relationships between proteins. It can generate novel proteins given partial sequence, structure, or functional constraints.

Limitations & risks

Validate outputs experimentally. The model may generate proteins that are not synthesizable or functional. Not intended for clinical or therapeutic applications without further validation.This model is released under the MIT License.


Parameters	Approximately 300 million, 600 million, and 6 billion parameters for ESMC 300M, ESMC 600M, and ESMC 6B respectively.
Transformer Layers	30, 36, and 80 layers for ESMC 300M, ESMC 600M, and ESMC 6B respectively.
Type of Model Architecture	Transformer architecture where each model consists of a token embedding layer, a stack of transformer blocks, and a regression output head. It features Pre-LN, rotary embeddings, and SwiGLU activations. No biases are used in linear layers or layer norms.
Training FLOPs	1×10²², 1×10²², and 2×10²³ for ESMC 300M, ESMC 600M, and ESMC 6B respectively.
Training Datasets	ESMC was trained on data pulled from UniRef, MGnify, and the Joint Genome Institute (JGI) Metagenome database.

Model	Model Size	Number of Layers	Max Context Length	Release Date
esmc-6b-2024-12	6B	80	16,384	2024-12
esmc-600m-2024-12	600M	36	16,384	2024-12
esmc-300m-2024-12	300M	30	16,384	2024-12


Primary Use Case	Protein representation learning and embeddings. Generating protein embeddings that can be fine-tuned for various downstream prediction tasks, including functional annotation, mutational effect analysis, and the design of novel proteins and peptides. Predicting the functional impact of mutations and amino acid substitutions on protein function.
Supported Input Modalities	Protein sequences
Access	Through Biohub or Hugging Face
License	This model is released under the MIT License.
Not Recommended For	Clinical diagnosis or treatment recommendations. Computational metrics do not replace wet-lab validation. Treat model outputs as machine-generated hypotheses that require further experimental validation, not as established biological facts.

ESM Atlas data

The ESM Atlas is a map of 6.8 billion proteins covering the full breadth of life’s biodiversity and more than one billion predicted structures. ESM Atlas organizes protein space by the shared representation space learned by ESMC, with high-resolution structures predicted by ESMFold2. Users can download the dataset used to generate the Atlas from AWS S3 at no cost. See the Atlas data files and download commands in the ESM Atlas guide.

Models