Insilico Medicine and Liquid AI Announce Strategic Partnership for Drug Discovery AI Models

Collaboration produces lightweight 2.6B-parameter foundation model that matches or outperforms larger systems across drug discovery benchmarks

Mar. 3, 2026 at 2:50pm

Got story updates? Submit your updates here. ›

Insilico Medicine and Liquid AI have announced a strategic partnership to develop lightweight scientific foundation models for pharmaceutical research. The collaboration has produced LFM2-2.6B-MMAI, a single 2.6 billion parameter model that achieves state-of-the-art performance across multiple drug discovery tasks, including property prediction, molecular optimization, affinity prediction, and chemical reasoning. The model can run entirely on private pharmaceutical infrastructure, addressing the challenge of harnessing cutting-edge AI capabilities without sending proprietary data to external cloud services.

Why it matters

This partnership demonstrates that efficient architecture design, not just scale, is key to making foundation models practical for scientific applications. By combining Liquid AI's efficient LFM architecture with Insilico's MMAI Gym training platform, the collaboration has produced a single model that can match or outperform much larger systems across the drug discovery pipeline, while operating entirely on private infrastructure. This unlocks immediate applications for pharmaceutical companies in areas like ADMET screening, lead optimization, and retrosynthesis assessment.

The details

The LFM2-2.6B-MMAI model covers the complete drug discovery loop, spanning property prediction, ADMET endpoints, multi-parameter molecular optimization, target-aware scoring, functional group reasoning, and retrosynthesis planning. Training involved approximately 120 billion tokens of pharmaceutical data across over two hundred different tasks. At just 2.6 billion parameters, the model outperformed a 27 billion parameter model on 13 of 22 property prediction tasks, reached up to 98.8% success rates on industry-standard molecular optimization benchmarks, and produced better affinity prediction results than frontier models like GPT-5.1 and Claude Opus 4.5.

  • The partnership was announced on March 3, 2026.

The players

Insilico Medicine

A clinical-stage biotechnology company using AI for drug development across cancer, fibrosis, immunity, central nervous system diseases, and aging-related conditions.

Liquid AI

A company that builds Liquid Foundation Models (LFMs) based on dynamical systems and signal processing, with a focus on efficient AI models that can be deployed on-premise or in resource-constrained environments.

MMAI Gym for Science

A domain-specific training environment developed by Insilico Medicine to elevate general-purpose and frontier Large Language Models (LLMs) into pharmaceutical-grade engines for drug discovery and development.

Ramin Hasani

CEO and co-founder of Liquid AI.

Alex Zhavoronkov

CEO of Insilico Medicine.

Got photos? Submit your photos here. ›

What they’re saying

“With LFM2-2.6B-MMAI, we've shown that efficient architecture design, not just scale, is what makes foundation models practical for the sciences. A single 2.6B-parameter model now matches or outperforms systems ten times its size across the drug discovery pipeline, all on private infrastructure. Our collaboration with Insilico is proof that you can reduce the cost of intelligence while raising the quality bar.”

— Ramin Hasani, CEO and co-founder of Liquid AI

“We are pleased to collaborate with Liquid AI to develop the next generation of lightweight liquid foundation models capable of performing multiple scientific tasks with state-of-the-art performance across drug discovery benchmarks. Highly-efficient liquid science models will make it easier for more scientists to achieve their goals in order to compress discovery timelines and ultimately help patients.”

— Alex Zhavoronkov, CEO of Insilico Medicine

What’s next

The LFM2-2.6B-MMAI model is now available for pharmaceutical companies to deploy on their private infrastructure and leverage across their drug discovery pipelines.

The takeaway

This partnership demonstrates that efficient AI architecture design, not just scale, is the key to making foundation models practical for scientific applications like drug discovery. By combining Liquid AI's LFM technology with Insilico's MMAI Gym, the collaboration has produced a lightweight model that can match or outperform much larger systems while operating entirely on private infrastructure, unlocking immediate benefits for pharmaceutical companies.