Rapid, structure-free binding affinity prediction enables OpenFold3 to address key bottleneck in commercial early-stage drug discovery

, /PRNewswire/ – SandboxAQ today launched AQAffinity, a powerful new AI model designed to unblock one of the most persistent bottlenecks in drug discovery: the need for experimentally determined structures. Built seamlessly on top of the OpenFold Consortium’s OpenFold3 biomolecular co-folding model, AQAffinity enables rapid, structure-free drug affinity prediction. By eliminating the dependency on pre-existing structures, AQAffinity allows researchers to "fail fast" on low-probability paths and double down on promising candidates earlier. This capability is critical for de-risking challenging targets—such as less structurally characterized proteins—effectively improving the odds of clinical success before wet lab cycles even begin.

"AI-driven binding affinity prediction has evolved from a niche method to a core component of modern drug discovery workflows. SandboxAQ is proud to deliver this powerful capability to the consortium and the open-source community," said Nadia Harhen, SandboxAQ General Manager for AI Simulation. "AQAffinity significantly enhances OpenFold3’s capabilities, giving it parity with competitive biomolecular co-folding models while adding new features and benefits that make it a more appealing choice for biopharma researchers and decision-makers."

OpenFold3 is available to all researchers and is fully transparent, unlike competitive models that are either unavailable for commercial use or challenging to evaluate due to opaque, "black box" results. AQAffinity provides all the data researchers need to evaluate model performance, including training data and methods, offering a more straightforward path to internal performance validation using their own benchmarks and easier fine-tuning or customization using their own data and relevant targets. Because OpenFold3 is fully open-source, organizations can evaluate the model, tools, and dataset for free, including experimenting on "hard targets" or ranking targets and ligands to decide where to invest wet lab cycles, with strong data security and no costly, long-term commitments to closed platforms.

"With AQAffinity, SandboxAQ delivered critical affinity prediction capabilities to OpenFold3 within a few short weeks after the model’s launch, showcasing the company’s technical expertise and commitment to advancing AI-powered open-source tools to accelerate drug discovery," said Professor Mohammed AlQuraishi, principal investigator at OpenFold. "The Consortium will continuously enhance OpenFold3’s performance and deliver an ecosystem of AI-powered drug discovery capabilities, including AQAffinity and other tools, to help members accelerate drug pipelines and bring novel therapies to market."

In December 2025, OpenFold Consortium members were allowed to beta test AQAffinity to understand its impact and compare its performance with other available tools.

"AQAffinity will complement OpenFold3 with valuable binding affinity predictions from SandboxAQ," said Sherry Liu, Co-founder and CTO at TamarindBio. "AQAffinity will add to Tamarind’s toolkit to help our customers streamline virtual screening, improve hit-to-lead optimization, and accelerate small molecule discovery and design."

AQAffinity is available now under the permissive Apache License 2.0 on Hugging Face. For more information on AQAffinity visit www.sandboxaq.com/aqaffinity or email [email protected].

**About SandboxAQ **‍SandboxAQ is a B2B company delivering solutions at the intersection of AI and quantum techniques. The company’s Large Quantitative Models (LQMs) deliver critical advances in life sciences, navigation, and other sectors. The company emerged from Alphabet Inc. as an independent, growth-backed company funded by leading investors and strategic partners including funds and accounts advised by T. Rowe Price Associates, Inc., Alger, IQT, US Innovative Technology Fund, S32, Paladin Capital, BNP Paribas, Eric Schmidt, Breyer Capital, Ray Dalio, Marc Benioff, Thomas Tull, Yann LeCun, and others. For more information, visit http://www.sandboxaq.com.

**About OpenFold Consortium **‍OpenFold is a non-profit 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. Membership is encouraged among Biotech, Pharma, Synthetic Bio, Software/Tech, and non-profit research organizations. For more information about OpenFold Consortium and its initiatives, please visit: https://openfold.io/

SOURCE SandboxAQ