SandboxAQ announced the general availability of AQPotency, a Large Quantitative Model that predicts drug-target activity without requiring a solved 3D structure. The model ranks molecule-target pairs in seconds and provides confidence scores, and has been used in eight customer programs with validated results.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
SandboxAQ announces the commercial availability of its large quantitative model AQPotency, which predicts drug-target activity without target 3D structures, at a cost of $1 per 1,000 comparisons. The model has been used in eight customer projects and claims experimental validation.
On August 19, 2026, SandboxAQ announced the commercial availability of its large quantitative model (LQM) AQPotency. The model is designed to predict the strength of interaction between candidate drug molecules and disease targets without requiring the target's 3D structure, breaking through the limitations of traditional computational methods. Traditional methods typically require detailed structural maps of the target, and many high-value targets lack such structures, making related projects difficult to advance.
AQPotency runs on standard computing hardware and can rank large molecular libraries in seconds, with costs as low as $1 per 1,000 comparisons. Additionally, the model provides confidence information for each prediction and determines whether the target is within the model's reliable range, making results more actionable.
Beyond forward prediction of drug-target activity, AQPotency also supports reverse screening: starting from a single promising molecule, it scans a broad panel of proteins in the body and returns a ranked list of potential targets. This helps research teams focus on possible targets for validation when a molecule shows useful effects but the mechanism is unknown.
This feature was mentioned by SandboxAQ's Vice President Andrea Bortolato in a statement, who said the model has been successfully used in eight customer projects with experimentally validated impact.
Professor Dario R. Alessi, Director of the MRC Protein Phosphorylation Unit at the University of Dundee, stated that SandboxAQ's model has significantly impacted his Parkinson's disease therapeutic development work, enabling them to explore a larger biochemical space in less time and improve activity and selectivity.
Dr. Gary W. Miller from Columbia University's Mailman School of Public Health mentioned that the collaboration with SandboxAQ combined AI discovery with rigorous experimental validation, identifying selective SV2C binders from commercial libraries, laying the foundation for potential new therapies for Parkinson's disease and other dopamine signaling disorders.
Robin Roehm, CEO of Apheris, believes that AQPotency makes high-value discovery decisions faster and more practical, allowing researchers to prioritize the most promising compounds with greater confidence and focus experimental resources on key areas.
AQPotency is now available via Claude's Model Context Protocol (MCP) and the SandboxAQ website, with Google Cloud Marketplace coming soon. Additionally, SandboxAQ also released another LQM, AQCat Adsorption Spin, for catalyst discovery, which is also officially available on Claude and Claude Science.
This report is based on SandboxAQ's press release and third-party industry media coverage, including company statements and customer testimonials. Customer testimonials may be subjective, and data such as 'eight customer projects' has not been independently verified and should be considered company claims rather than confirmed facts.
The launch of AQPotency marks progress in the field of AI-driven drug discovery: target screening without 3D structures, with low cost and high-confidence outputs to support decision-making. However, its actual effectiveness still requires more independent validation.
Primary report
Primary source