Insilico Medicine has launched Nach01, a new foundation model designed for generative chemistry, now available to organizations globally via the AWS Marketplace, the company said.
The model was developed using Amazon SageMaker and blends a large language model with spatial chemical data processing through point cloud transformers. The company claims this multimodal approach enhances tasks including retrosynthesis, molecular property prediction, and novel compound generation.
Trained on AWS SageMaker, Nach01 supports both inference and fine-tuning via Python APIs or cloud-native deployments. Insilico Medicine said users can implement the model directly within AWS environments, leveraging scalable infrastructure across a broad array of chemistry applications.
The company’s CEO, Dr. Alex Zhavoronkov, called Nach01 “one of the most capable multimodal foundational models in our arsenal and in the industry” and described it as a “stepping stone on our path to pharmaceutical superIntelligence,” the company said.
Jon Jones, AWS vice president and global head of startups, said the launch highlights AWS’s mission “to make frontier biochemistry models like NACH01‑PC broadly available to the research community around the world,” helping to speed the discovery of improved treatments and enhance patient outcomes, the company said.
Nach01 joins the Nach0 family of foundation models and is intended for use in drug discovery workflows spanning hit finding, lead optimization, and property modeling. Organizations worldwide may request access or purchase the model through AWS Marketplace, according to the company.
