The OpenFold Consortium today announced a major OpenFold3 update and the public release of training datasets and full-stack tooling for reproducible biomolecular AI. OpenFold3 is an open-source deep l...
Autore: Business Wire
BERKELEY, Calif.: The OpenFold Consortium today announced a major OpenFold3 update and the public release of training datasets and full-stack tooling for reproducible biomolecular AI. OpenFold3 is an open-source deep learning system for cofolding that predicts the 3D structures of biomolecular complexes from sequence and molecular inputs, including proteins interacting with small molecules and nucleic acids.
“OpenFold was built on the principle that foundational AI for biology should be open, reproducible, and auditable,” said Woody Sherman, Executive Committee Chairperson of the OpenFold Consortium and Chief Innovation Officer at PsiThera. “By releasing OpenFold3 with open data, permissive licensing, and transparent workflows, we are enabling independent validation and rapid iteration so researchers can turn cofolding models into reliable scientific infrastructure that accelerates drug discovery and deepens our mechanistic understanding of biology.”
OpenFold3 enables structure prediction for biomolecular complexes relevant to drug discovery, protein engineering, and basic research, supporting both evaluation workflows and downstream method development.
“Releasing the full training stack behind OpenFold3 is a major milestone for reproducible biomolecular AI. It allows the community not only to run the model, but to inspect, retrain, and push the technology forward,” said Nazim Bouatta, Ph.D, OpenFold Advisor.
With this update, OpenFold3 is available as an end-to-end open cofolding stack, including training datasets, model weights, training and inference code, and evaluation scripts released under permissive licenses. This full-stack release enables independent reproduction of reported results, rigorous benchmarking, and extension through fine-tuning and method development-capabilities that are difficult to achieve with closed or inference-only systems.
Updated benchmarks in the OpenFold3 white paper report competitive performance versus AlphaFold3 across a broad set of modalities and evaluation tasks, reinforcing that open development can deliver state-of-the-art results while preserving transparency, extensibility, and independent validation.
Open Data for Reproducible Biomolecular AI
As part of this update, the OpenFold Consortium is releasing OpenFold3 training datasets via the Registry of Open Data on AWS, lowering the barrier for independent training, benchmarking, and method comparison. This dataset release is intended to support reproducible research by enabling teams to validate results, retrain models for new scientific questions, and compare new approaches on shared foundations.
This release builds on the OpenFold Consortium’s prior OpenFold3 preview and reflects substantial progress in model quality, usability, and infrastructure. Updated benchmarks, evaluation resources, code, and dataset documentation are available via the OpenFold3 portal and the public OpenFold3 repository on GitHub.
A Model Built for the Community
To support adoption, the consortium is launching the OpenFold3 portal, which provides onboarding materials, installation and deployment guidance, reference inference pipelines, evaluation scripts, and dataset access documentation, supported by a public community channel for technical Q&A and issue triage.
These resources are designed to help developers and research teams move quickly from evaluation to deployment, while enabling feedback and contributions that strengthen the ecosystem over time. By combining open infrastructure with active community support, the consortium aims to accelerate both adoption and collective improvement of the system.
“The power of OpenFold3 is that it’s not just a model, it’s a fully open foundation the community can adapt and extend,” said Arman Zaribafiyan, Head of Strategic Alliances at SandboxAQ. “At SandboxAQ, we’ve already built on earlier OpenFold advances in our AQAffinity model for binding‑affinity prediction, and we’re now adopting OpenFold3 to supercharge those capabilities even further. That kind of open, collaborative development is exactly what biomolecular AI needs to deliver meaningful results, accelerated discovery and eventually new medicines faster.”
What Comes Next
The consortium noted that antibody–antigen complex prediction remains a challenging frontier for the field, with current methods (including OpenFold3) showing room for improvement in this domain. Improving antibody–antigen performance is a major 2026 priority, with planned work spanning data expansion, benchmarking, and model improvements targeted to immune-relevant complexes.
By pairing competitive performance across most evaluated modalities with full-stack openness, OpenFold3 provides a platform the community can use today while helping define a rigorous, benchmark-driven agenda for what comes next.
A Shared Foundation for Future Innovation
The OpenFold Consortium was created to ensure that foundational AI infrastructure for biology remains open to the global scientific community. With this OpenFold3 update and dataset release, the consortium is extending its mission into the cofolding era by providing an openly available platform that researchers can deploy, audit, retrain, and build upon.
As biological AI becomes increasingly central to drug discovery and molecular design, the consortium believes that open systems will be essential to ensuring scientific reproducibility, broad access, and rapid innovation across academia, biotech, pharma, and nonprofit research. Sustaining open, reproducible foundation models at this scale requires ongoing investment in compute, dataset generation, and research software engineering, and the consortium welcomes partners who want to support this shared infrastructure.
About OpenFold
OpenFold is a nonprofit AI research consortium of academic and industry partners whose goal is to develop free and open-source software tools for biology and drug discovery, hosted as a project of the Open Molecular Software Foundation (OMSF). Membership is open to organizations across biotech, pharma, synthetic biology, software/technology, academia, and nonprofit research.
For more information, please visit OpenFold’s website.
Fonte: Business Wire