LAGUNA HILLS, California & SHANGHAI, China, July 6, 2026
Aureka, an AI TechBio company developing next-generation infrastructure for AI-driven biologics discovery, has announced the release of Open Drug Discovery Engine (OpenDDE), an open-source, all-atom biomolecular foundation model designed to accelerate AI-powered therapeutic discovery. Released under the Apache-2.0 open-source license, OpenDDE is intended to serve as the structural reasoning core for future drug discovery systems by modeling complex interactions between proteins, nucleic acids, antibodies, small molecules, and other biomolecular components. Rather than functioning solely as a protein structure prediction model, OpenDDE introduces an integrated platform capable of supporting sequence-structure-function modeling, laying the groundwork for future applications in de novo drug design, affinity prediction, structure-guided optimization, and closed-loop therapeutic discovery workflows. By combining cutting-edge artificial intelligence with scalable biomolecular modeling, Aureka aims to democratize access to advanced drug discovery technologies while enabling researchers, biotechnology companies, and academic institutions worldwide to accelerate biomedical innovation.
OpenDDE Introduces Next-Generation AI Foundation for Drug Discovery
At the core of OpenDDE is a sophisticated all-atom biomolecular foundation model that uses co-folding as the entry point for understanding interactions across multiple biological molecules simultaneously. Unlike traditional structure prediction tools that focus primarily on individual proteins, OpenDDE integrates atomic-level latent reasoning, allowing the model to refine representations of molecular geometry, chemical context, and complex biomolecular interfaces before generating highly detailed structural predictions. The platform has been specifically engineered to become the foundation of an extensible AI drug discovery engine, supporting future capabilities including antibody engineering, molecular design, binding affinity estimation, conformational modeling, and experimental optimization.
According to Aureka, the model demonstrated competitive antibody-antigen co-folding performance across multiple benchmark datasets while narrowing the performance gap with leading proprietary systems. These capabilities position OpenDDE as a significant advancement in applying artificial intelligence to structural biology, enabling more accurate prediction of molecular interactions that underpin therapeutic discovery.
Large-Scale AI Infrastructure Supports Biomedical Innovation
Advertisement
OpenDDE reflects the increasing scale of artificial intelligence infrastructure required for modern biomedical research. The model contains approximately 655 million trainable parameters and required nearly 414,000 GPU-hours for training, highlighting the growing computational demands of biomolecular foundation models. Aureka’s research also identified clear scaling laws demonstrating that larger datasets, more powerful models, enhanced inference techniques, and improved training strategies consistently strengthen biological reasoning and structural prediction performance. Beyond computational modeling, Aureka is integrating OpenDDE with a high-throughput automated wet-laboratory platform, creating a closed-loop dry-lab and wet-lab discovery system capable of continuously improving AI-generated therapeutic candidates through experimental validation.
This integrated platform combines autonomous antibody design, single-cell functional screening, and automated yeast evolution, allowing AI systems to generate, test, analyze, and refine antibody candidates through iterative experimental feedback. Such infrastructure is expected to accelerate discovery across challenging therapeutic areas including multispecific antibodies, epitope-specific antibodies, internalizing antibodies, and pH-switch antibodies.
Open-Source Platform Expands Global AI Drug Discovery Collaboration
By releasing OpenDDE as a fully open-source platform, Aureka aims to make advanced biomolecular AI accessible to researchers worldwide while promoting scientific collaboration across academia, biotechnology, and the pharmaceutical industry. The release includes training code, inference pipelines, pretrained model checkpoints, and benchmark datasets, enabling independent validation and continued community-driven development. Although the current version focuses primarily on biomolecular structure prediction and antibody-antigen modeling, Aureka plans to expand the platform to support de novo molecular design, affinity prediction, structure-conditioned optimization, experimental learning, and broader AI-driven therapeutic discovery applications.
The company’s long-term vision centers on building comprehensive TechBio infrastructure that combines artificial intelligence, automation, high-performance computing, and experimental biology to accelerate the development of first-in-class and best-in-class biologic therapies. As AI continues transforming biomedical research, OpenDDE represents an important milestone toward creating scalable, transparent, and collaborative platforms capable of accelerating innovation across the global drug discovery ecosystem.
Source: Aureka press release
