SandboxAQ Integrates its Quantitative AI Models with Anthropic's Claude via MCP

SandboxAQ today announced the integration of its Large Quantitative Models (LQMs) with Claude, Anthropic's frontier AI model, making it possible to directly connect a large language model to a large quantitative model for drug discovery and materials science. Trained on real-world lab data and scientific equations, LQMs are AI models engineered for the quantitative economy, a $50+ trillion sector spanning biopharma, financial services, energy, and advanced materials.

Until now, running the advanced models for new drug discovery and materials discovery required specialized scientists and the ability to write complex code. With Claude serving as a natural language interface to SandboxAQ's LQM platform, any user can access that same capability through natural language prompts, moving faster from hypothesis to discovery, in the physical world. 

Built on Physics and the Real World

SandboxAQ's LQMs are accelerating drug discovery and materials science, with active programs underway at major pharmaceutical companies and demonstrated advances in battery chemistry, catalysts, and alloys. A number of SandboxAQ’s frontier models, including AQAffinity, have been built on OpenFold and developed in collaboration with NVIDIA.

SandboxAQ’s platform also powers AI-native cybersecurity, medical research, and navigation systems, with financial services and risk modeling modules going live soon. 

SandboxAQ builds its proprietary LQMs from the ground up, generating its own physics-grounded training data through high-fidelity simulations, including quantum chemistry calculations, molecular dynamics, and microkinetics, targeting the specific chemistries and conditions that matter most. SandboxAQ can augment these data sets with data from lab experiments. SandboxAQ trains its own AI models on that data, owns them outright, and connects them into automated workflows that run full design, test, and decision cycles, allowing users to move from question to defensible answer without writing a single line of code.

AQCat, Now Accessible via Claude, Accelerates the Most Critical Step in Catalyst Discovery

As of today's integration with Claude, users can access AQCat Adsorption Spin. AQCat allows users to lock in the most critical first step of any catalyst discovery workflow, adsorption energy calculation (a measure of how strongly molecules bind to a catalyst surface), allowing them to rapidly identify and prioritize the most promising candidates before committing costly modeling and lab resources to full-scale evaluation. AQCat Adsorption Spin gives users gold-standard accuracy at a fraction of the time and cost, unlocking materials screening at a scale that was previously out of reach.

Catalysts underpin more than 90% of all commercially produced chemical products, and the ability to screen at unprecedented speed and accuracy has direct impact across green hydrogen, sustainable aviation fuel, fertilizer production, plastics recycling, and more.

"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,” said 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.

"Now, researchers can access frontier physics-based models directly inside the AI tools they already use, with no additional infrastructure, code or barriers," said Jack D. Hidary, CEO of SandboxAQ. "Our LQMs bring the rigor of first-principles quantum chemistry to a conversational interface, and that changes how fast a user can move from question to answer across chemistry, materials science, and drug development."

Drug Discovery Models Coming Soon to Claude

SandboxAQ's work with Claude extends beyond catalysis. A suite of drug discovery models will soon be accessible through the same natural language interface, bringing SandboxAQ's pharmaceutical AI capabilities to a broader range of R&D teams. Models coming to the platform include:

• AQPotency, which will allow users to identify and prioritize the most promising drug candidates computationally, screening thousands of options at a fraction of the time and cost of traditional methods.
• AQCell, which will enable users to simulate how living cells respond to drug candidates across thousands of compounds, predicting whether a drug activates the right biological pathway and flagging potential liver toxicity.

“Connecting physics-grounded quantitative models from SandboxAQ with large language models like Claude removes a critical barrier between researchers and the frontier of computational science,” said Woody Sherman, PhD, Chief Innovation Officer and Founder, PsiThera, and Executive Committee Chairperson, OpenFold Consortium.

Robin Röhm, CEO and Co-Founder of Apheris, said: “Bringing quantitative AI into the tools pharma teams already use is the kind of shift that can fundamentally accelerate the pace of drug discovery."

"As we bring these capabilities to Claude, users in pharma and biotech will be able to run workflows that previously required weeks of computational setup in hours," said Nadia Harhen, General Manager of AI Simulation at SandboxAQ. "Our drug discovery models are built on the same physics-grounded infrastructure that powers AQCat Adsorption Spin today. This means any researcher, regardless of their technical background, can find a faster path from scientific question to answer.”

Users interested in accessing SandboxAQ's LQMs through Claude can join the waitlist here. Additional models and integrations are coming soon.­