SandboxAQ’s Models Link Large Language AI to $50T Economy

SandboxAQ is connecting its Large Quantitative Models (LQMs) directly to Anthropic’s Claude, expanding access to advanced AI for scientific discovery. These LQMs, engineered for sectors like biopharma and energy with a combined economic impact exceeding $50 trillion, were previously limited to users with specialized coding skills. The integration allows researchers to utilize plain-English prompts to run complex simulations for areas such as drug discovery and materials science, accelerating the process “from hypothesis to discovery, in the physical world.” Several of SandboxAQ’s models, including AQAffinity and AQCat, were developed in collaboration with NVIDIA. Partha P. Mukherjee, Ph.D., Professor & University Faculty Scholar, School of Mechanical Engineering, and Director, Center for Advances in Resilient Energy Storage (CARES), Purdue University.

Large Quantitative Models Integrated with Anthropic’s Claude

The convergence of artificial intelligence and quantitative modeling is expanding the scope of scientific discovery, as demonstrated by SandboxAQ’s integration of its Large Quantitative Models (LQMs) with Anthropic’s Claude. This connection unlocks access to sophisticated AI previously limited by the need for specialized expertise and extensive coding skills; researchers can now leverage these tools with simple, plain-English prompts. SandboxAQ’s models are designed to serve a quantitative economy exceeding $50 trillion, encompassing crucial sectors like biopharmaceuticals, finance, and materials science. A key component of this advancement is the development of models like AQAffinity and AQCat, developed in collaboration with NVIDIA. Currently, users can access AQCat Adsorption Spin through Claude, accelerating the critical initial step in catalyst discovery, calculating adsorption energy, at a reduced cost and timeframe.

Catalysts are fundamental to over 90% of commercial chemical products, so improvements in their discovery have broad implications for industries ranging from green hydrogen to plastics recycling. Partha P. Mukherjee, Ph.D., Professor & University Faculty Scholar, School of Mechanical Engineering, and Director, Center for Advances in Resilient Energy Storage (CARES), Purdue University, said, “SandboxAQ’s integration with Claude removes one of the key barriers between a researcher’s scientific intuition and rigorous physics-grounded computation, accelerating discovery across energy materials and beyond.” SandboxAQ plans to release a suite of drug discovery models accessible through Claude, including AQPotency and AQCell, promising to expedite pharmaceutical research and development with physics-grounded AI.

AQCat Adsorption Spin Accelerates Catalyst Discovery

The pursuit of novel catalysts, substances that accelerate chemical reactions, has long relied on a blend of intuition, experimentation, and increasingly, computational modeling. Accessing and utilizing these advanced models previously demanded expertise in scientific coding and specialized knowledge, creating a bottleneck in materials discovery. SandboxAQ is addressing this challenge with AQAffinity and AQCat, Large Quantitative Models (LQMs) designed to integrate directly with Anthropic’s Claude, a frontier AI model. These LQMs, trained on extensive real-world laboratory data and fundamental scientific equations, serve a quantitative economy estimated at over $50 trillion encompassing sectors from biopharmaceuticals to energy. A particularly notable advancement is AQCat Adsorption Spin, now accessible through Claude, which focuses on the crucial initial step of catalyst discovery: adsorption energy calculation. This calculation determines how strongly molecules bind to a catalyst’s surface, allowing researchers to quickly prioritize promising candidates before investing in resource-intensive modeling and laboratory work.

According to SandboxAQ, AQCat delivers accuracy at a fraction of the time and cost, enabling materials screening at an unprecedented scale with impacts spanning green hydrogen, sustainable aviation fuel, and plastics recycling. Partha P. Mukherjee, Ph.D., Professor & University Faculty Scholar, School of Mechanical Engineering, and Director, Center for Advances in Resilient Energy Storage (CARES), Purdue University, further explained, “Now, researchers can access frontier physics-based models directly inside the AI tools they already use, with no additional infrastructure, code or barriers.”

Physics-Grounded Models for Drug Candidate Screening

SandboxAQ is expanding access to its sophisticated physics-grounded AI models, initially through integration with Anthropic’s Claude, and altering workflows within the $50+ trillion quantitative economy. Previously, utilizing these Large Quantitative Models (LQMs) for applications like drug discovery and materials science demanded specialized expertise in coding and scientific computation; now, users can leverage plain-English prompts to access the same capabilities. AQCat allows researchers to rapidly prioritize promising catalyst candidates, reducing the time and cost associated with extensive modeling and laboratory evaluation, a capability critical given that catalysts underpin over 90% of all commercially produced chemical products. SandboxAQ builds its LQMs using high-fidelity simulations, including quantum chemistry calculations, and augments this with experimental data, ensuring ownership and control over the models and the automated workflows they power. Partha P. Mukherjee, Ph.D., Professor & University Faculty Scholar, School of Mechanical Engineering, and Director, Center for Advances in Resilient Energy Storage (CARES), Purdue University, emphasized the platform’s potential to transform how quickly users can move from scientific question to defensible answer.

SandboxAQ’s LQMs Power Diverse Quantitative Applications

Through a natural language interface powered by Claude, users can now access the capabilities of SandboxAQ’s platform using plain-English prompts, streamlining the process from initial hypothesis to tangible results. This advancement is particularly evident in the accessibility of AQCat Adsorption Spin, a model that accelerates catalyst discovery by rapidly calculating adsorption energies, a critical first step often requiring significant computational resources. Partha P. Mukherjee, Ph.D., Professor & University Faculty Scholar, School of Mechanical Engineering, and Director, Center for Advances in Resilient Energy Storage (CARES), Purdue University, said, “SandboxAQ’s integration with Claude removes one of the key barriers between a researcher’s scientific intuition and rigorous physics-grounded computation, accelerating discovery across energy materials and beyond.” Jack D. Hidary, CEO of SandboxAQ, noted that the company builds its models from the ground up, generating its own training data through high-fidelity simulations, ensuring ownership and control over the entire workflow.