CAMBRIDGE, Mass., April 14, 2026 — Insilico Medicine (“Insilico”, 3696.HK), a clinical-stage generative artificial intelligence (AI)-driven biotechnology company today announced the expansion of its MMAI Gym, a foundation-model training framework and platform for large-scale training and benchmarking, with the launch of three benchmark leaderboard portals to evaluate and improve AI systems across scientific research and drug discovery.
Positioned as both a “trainer and benchmark” for scientific AI, MMAI Gym enables organizations to train models for domain-specific reasoning while rigorously assessing their performance across real-world tasks.
MMAI Gym includes diverse benchmark tasks spanning drug discovery, biology, chemistry, and broader scientific domains. From this system, Insilico is launching benchmark leaderboard portals structured around three flagship categories:
ScienceAI Bench: Assesses broader scientific reasoning across biology, chemistry, longevity, materials science, and agriculture ( scienceaibench.insilico.com )
Drug Discovery Benchmark (DDB): Evaluates end-to-end drug discovery tasks, including target identification, molecular design, and optimization ( ddb.insilico.com )
Insilico Bench: Proprietary benchmarks developed by Insilico for drug discovery and other complex scientific challenges ( insilicobench.insilico.com ). For example, TargetBench, designed to assess target identification capability, has been used to validate the performance of Insilico’s latest target discovery model TargetPro, with related study has been published in Scientific Reports .
The platform combines curated industry-standard benchmarks with proprietary datasets, many grounded in experimental data, to ensure performance reflects real-world utility. Benchmark categories are intentionally overlapping to enable multi-dimensional evaluation across both generalist and specialized tasks.
To support transparency and adoption, Insilico is launching public leaderboard portals across all three categories, initially covering more than 200 benchmark tasks, with further expansion planned.
“MMAI Gym creates a unified system to train, evaluate, and continuously improve AI for science,” said Alex Zhavoronkov, CEO of Insilico Medicine. “By establishing standardized benchmarks and training environments, we are enabling scalable, trustworthy AI for drug discovery.”
MMAI Gym represents a shift toward standardized, scalable evaluation of scientific AI, providing a common framework for training, benchmarking and comparing models while accelerating their adoption across pharmaceutical R&D.
Previously, Insilico demonstrated that MMAI-trained foundation models achieved up to 10X performance gains on key drug discovery benchmarks compared to general-purpose foundation models, which fell short on approximately 75–95% of tasks. Moreover, in March 2026, Insilico and Liquid AI jointly delivered LFM2-2.6B-MMAI (v0.2.1), the first model trained through their first MMAI Gym collaboration. Despite its lightweight, on-premise design, the model delivered SOTA performance across several key tasks. The paper detailing the training process and final performance was accepted at ICLR 2026.
About Insilico Medicine
Insilico Medicine is a pioneering global biotechnology company dedicated to integrating artificial intelligence and automation technologies to accelerate drug discovery, drive innovation in the life sciences, and extend healthy longevity to people on the planet. The company was listed on the Main Board of the Hong Kong Stock Exchange on December 30, 2025, under the stock code 03696.HK.
By integrating AI and automation technologies and deep in-house drug discovery capabilities, Insilico is delivering innovative drug solutions for unmet needs including fibrosis, oncology, immunology, pain, and obesity and metabolic disorders. Additionally, Insilico extends the reach of Pharma.AI across diverse industries, such as advanced materials, agriculture, nutritional products and veterinary medicine. For more information, please visit www.insilico.com