As the AI era is increasingly shaped by foundational models, the pharmaceutical industry is entering a new phase of discovery, design, and decision-making opportunities driven by scientific AI. To explore these advances, Insilico Medicine (03696.HK), a clinical-stage generative AI-driven drug discovery company, today announced that Pharma.AI Spring Kickoff 2026 will be held on April 14th at 10am ET. Registration and event details are available below. https://insilico.zoom.us/webinar/register/WN_h7tujok6SdmfDWzkZwRgNg.
The 2026 season of the Pharma.AI Webinar Series will introduce you to the ongoing AI revolution in life sciences, including the growing interest in using fundamental models and why specialized models remain essential in biology, chemistry, and translational research. How Pharma.AI integrates fundamental models and scientific AI agents within an AI-driven integrated workflow for pharmaceutical R&D and scientific research. We will also discuss how Insilico’s cutting-edge ‘AI training AI’ approach can further adapt fundamental models to scientific and drug discovery applications and accelerate the evolution of AI decision-making systems.
More specifically, upcoming events will focus on new features across the Pharma.AI ecosystem, including updates to core modules such as MMAI Gym for Science, PandaOmics, Generative Biologics, and Chemistry42.
As we begin 2026, we are focused on moving beyond simple AI-driven to a true AI decision-making ecosystem. With the introduction of the continuously evolving Pharma.AI, we are building the foundation for a pharmaceutical superintelligence system that can reason more effectively, adapt to real-world scientific workflows, and create meaningful impact across drug discovery and development. Upcoming webinars bring together exciting new updates and are designed to provide researchers with the latest tools and best practices to tackle the most challenging problems in human health. ”
Dr. Alex Alipar, President, Insilico Medicine
Highlights at a glance
- MMAI Gym: Turning basic models into high-performance drug discovery engines
The MMAI Gym for Science, a foundational model training framework, was introduced by Insilico in January 2026. The framework leverages over 1,000 drug R&D benchmarks and approximately 120 billion tokens of public and proprietary drug discovery data, and leverages multi-tasking fine-tuning and reinforcement learning to significantly improve the performance of underlying models across specialized tasks in drug discovery.
Validating the capabilities of this framework, we demonstrated that a base model trained on MMAI achieved up to 10x performance improvement on key drug discovery benchmarks compared to a generic base model that was inadequate in approximately 75-95% of tasks. Additionally, in March 2026, Insilico and Liquid AI jointly delivered LFM2-2.6B-MMAI (v0.2.1), the first model trained through the first MMAI Gym collaboration. This model achieved SOTA performance across several critical tasks despite its lightweight on-premises design. A paper detailing the training process and final performance was accepted at ICLR 2026.
At the upcoming event, attendees will learn how this supervised fine-tuning (SFT) and reinforced fine-tuning (RFT) training and benchmarking system can significantly improve the performance of causal LLMs in real-world drug discovery tasks, and how to access the platform.
- PandaOmics: Target prioritization with single cells and panda claw
PandaOmics is Insilico Medicine’s AI-driven platform for therapeutic target discovery and indication expansion. Integrating and analyzing large multi-omics and biomedical datasets helps researchers identify and prioritize disease-specific drug targets and expand therapeutic indications for targets of interest.
Recent upgrades to PandaOmics incorporate comprehensive single-cell datasets that improve the resolution of target identification. In addition, panda clawis an agent AI tool that enables scientists to conduct complex real-time multi-omics analyses, generate research hypotheses, and perform target evaluations through a simple natural language interface.
- Chemistry42: Multiple targets and advanced alchemy
Chemistry42 is Insilico Medicine’s AI-driven platform for designing and discovering new small molecules. Combining generative model ensembles with advanced physics-based techniques to help researchers create and optimize new compounds. The core of Chemistry42 is Nach01. Nach01 is an AI model trained on billions of data points to understand both natural and chemical language, enabling hundreds of specialized tasks and laying the foundation for the future of “quick medicine.”
Latest updates include multi-target support for molecule generation, enhanced result visualization for smoother analysis, Nach01-MMAI for molecule generation, and new absolute binding free energy (ABFE) calculations in Alchemistry.
- Generative biologics: cyclic peptide design and linear peptide optimization
Generative Biologics is a cutting-edge biologics engineering platform. Tackle complex challenges in the design of antibodies, peptides, and other biologics using advanced multiparameter optimization. Generative Biologics leverages more than 10 generative and predictive models and is powered by accurate physics-based tools to enable rapid creation of diverse and optimized biologics, enabling scientists to generate viable binder candidates within 72 hours.
This platform includes major updates for peptide design. Introducing a completely new workflow. cyclic peptidesupports both head-to-tail and disulfide bond architectures, and generates hundreds of candidates in just a few hours with AI and physics-based prioritization. In parallel, the researchers successfully used the platform to optimize a linear peptide and purified lead candidate P3 against GLP-1R, generating dozens of new candidates, with the top variant, P3-1, achieving a 6-fold improvement compared to the original lead.
Pharma.AI is an end-to-end AI platform for drug discovery and development that integrates target discovery, generative chemistry, biologics design, and predictive clinical modeling into an integrated AI-driven workflow for drug R&D. As we kick off 2026, we look forward to seeing you at our first event.
date: April 14, 2026
time: 10am (Eastern Standard Time)
