Boltz-1

Pre-trained model for all-atom structure prediction of biomolecular complexes.

Disclaimer

This is an UNOFFICIAL implementation of Boltz-1: Democratizing Biomolecular Interaction Modeling by Jeremy Wohlwend, Gabriele Corso, et al.

The OFFICIAL repository of Boltz-1 is at jwohlwend/boltz.

The team releasing Boltz-1 did not write this model card for this model so this model card has been written by the OFoldX team.

Model Details

Boltz-1 refines atom and token features with an MSA module and a recycled Pairformer trunk. An EDM diffusion model generates coordinates, and a dedicated confidence trunk predicts pLDDT, PAE, PDE, pTM, ipTM, and ranking scores. The model also supports pocket conditioning and Boltz-1x sampling-time steering.

Please refer to the Training Details section for more information on the training process.

Variants

oteam/boltz1: Biomolecular structure and confidence prediction.

Boltz-1x physical-potential and Feynman-Kac steering are inference controls on this model, rather than separate weights.

Model Specification

Pairformer Blocks	MSA Blocks	Denoiser Blocks	Confidence Blocks	Single Size	Pair Size	Trunk Passes	Sampling Steps	Num Parameters (M)
48	4	24	48	384	128	4	200	592.163

Usage

The model file depends on the ofoldx library. You can install it using pip:

pip install ofoldx

Direct Use

Structure Prediction

You can use this model directly with a pipeline for all-atom structure prediction:

import ofoldx
from ofoldx.io import read_structure

structure, report = read_structure("input.cif")
pipeline = ofoldx.pipeline("fold", model="oteam/boltz1")
result = pipeline(structure, seed=7)

The result contains predicted structures and confidence values.

Downstream Use

Physical Guidance

Boltz-1x supports physical-potential and Feynman–Kac guidance during sampling. Set use_potentials=True in the pipeline call to enable it.

Training Details

Boltz-1 is trained for biomolecular structure prediction with coordinate denoising and confidence objectives.

Training Data

The Boltz-1 structure model uses PDB structures and OpenFold self-distillation data. The upstream project provides the preprocessed data and processing pipeline.

Training Procedure

Boltz-1 uses EDM coordinate denoising together with distogram, confidence, and ranking objectives.

Citation

@article{wohlwend2024boltz1,
  author = {Wohlwend, Jeremy and Corso, Gabriele and Passaro, Saro and others},
  title = {Boltz-1: Democratizing Biomolecular Interaction Modeling},
  year = {2024},
  doi = {10.1101/2024.11.19.624167},
  journal = {bioRxiv}
}

The artifacts distributed in this repository are part of the OFoldX project. If OFoldX supports your research, please cite the OFoldX project as follows:

@software{chen2026ofoldx,
  author = {Chen, Zhiyuan},
  title  = {OFoldX},
  year   = {2026},
  url    = {https://github.com/OTeam-AI4S/OFoldX},
}

Contact

Please use GitHub issues of OFoldX for any questions or comments on the model implementation or model card.

Please contact the authors of the Boltz-1 project for questions or comments on the paper/model.

License

The model weights, implementation, configuration, and documentation distributed by OFoldX are licensed under the GNU Affero General Public License.

For additional terms and clarifications, please refer to our License FAQ.

SPDX-License-Identifier: AGPL-3.0-or-later

Downloads last month: -; Downloads are not tracked for this model. How to track

Safetensors

Model size

0.6B params

Tensor type

F32

Inference Providers NEW

Other

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

oteam
/

boltz1