Insilico Medicine and Liquid AI launch a secure, lightweight AI to streamline drug research and development.
Discovering new medicines doesn’t have to take years of trial and error. Promising compounds aren’t lost in a mountain of experiments, and patient breakthroughs happen faster. That’s the vision behind the new partnership between Insilico Medicine and Liquid AI, two companies aiming to make AI not just powerful, but practical for the pharmaceutical world [1].
The collaboration has produced LFM2-2.6B-MMAI, a single AI model that can tackle multiple drug discovery tasks – from predicting how molecules behave in the body to designing new compounds – all while running entirely on private company servers. Made for chemists, packed with intelligence but light enough to fit in their lab’s own “digital toolbox.”
Most AI models used in drug development are enormous and require cloud supercomputers and large amounts of data. This new model takes a different approach. At just 2.6 billion parameters, it achieves results that match or even outperform models ten times its size, all without sending sensitive information to the cloud.
Ramin Hasani, the CEO of Liquid AI, explained that the practicality of foundation models in the sciences stems from their efficient architecture rather than simply from increased scale. He noted that their partnership with Insilico serves as evidence that it is possible to lower the cost of intelligence while simultaneously elevating the quality of the output.
For companies, this is a game-changer. Proprietary molecular data, clinical targets and other sensitive information can stay on-premise, reducing risk while letting scientists leverage the full power of AI.
What does this AI actually do? In human terms, it helps researchers answer questions like:
- Which molecules are likely safe and effective?
- How can a compound be tweaked to work better across multiple criteria?
- Which molecules will stick to the right protein targets?
- How can a molecule be made in the lab efficiently?
The model even achieved up to 98.8% success in multi-parameter optimization, a kind of AI “molecular coaching” that helps chemists hit multiple goals at once. It also excels in chemical reasoning, suggesting logical synthetic steps for new compounds.
Why does longevity research stand to benefit? Insilico Medicine has a clear focus: aging-related diseases. From fibrosis to neurodegenerative disorders, the company’s AI platform aims to shorten the timeline from discovery to treatment. By streamlining early-stage drug discovery, models like LFM2-2.6B-MMAI could help therapies for age-related conditions reach patients sooner.
Alex Zhavoronkov, CEO of Insilico Medicine, frames it plainly: “Highly-efficient liquid science models will make it easier for more scientists to achieve their goals, compress discovery timelines, and ultimately help patients.”
For the longevity sector, faster discovery is both convenient and transformative. Every saved month in the lab could translate to earlier interventions that help people live longer, healthier lives.
From an investment perspective, this partnership highlights a subtle but critical shift in pharma AI. Instead of chasing ever-larger cloud models, Insilico and Liquid AI are betting on smaller, nimble systems that offer top-tier performance and data privacy. For large pharmaceutical companies, that’s a compelling combination of speed, efficiency and confidentiality.
If widely adopted, this model could create new enterprise revenue streams for Insilico Medicine, while strengthening its position in AI-driven drug discovery. The emphasis on on-premise, secure deployment sets it apart in a crowded market where data governance and regulatory compliance are increasingly non-negotiable.
This partnership goes beyond making AI “smarter.” It’s rethinking the drug discovery process itself, making it faster, safer and more accessible to researchers. For longevity, that could mean earlier breakthroughs, better-targeted therapies and a future where age-related diseases are tackled before they become life-limiting.
As AI becomes increasingly integral to pharmaceutical innovation, Insilico Medicine and Liquid AI are showing that efficiency, security and versatility can be just as important as raw computational power. Small, focused and smart – that may well be the future of discovering new medicines.
[1] https://www.liquid.ai/press/liquid-ai-insilico-medicine-partnership
