Overview

• EXAONE Path 2.0 is trained on over 10,000 paired whole-slide images and multiomics datasets to predict gene activity without genomic testing
• The model achieves state-of-the-art 78.4 percent accuracy in mutation prediction, surpassing competing systems
• Genetic test times have been cut from over two weeks to under one minute, enabling faster therapeutic decisions for cancer patients
• LG AI Research has partnered with Vanderbilt University Medical Center to develop a multimodal medical AI platform for clinical deployment
• Specialized AI models now target lung and colorectal cancers, and collaborations with Jackson Laboratory and Seoul National University extend its reach into Alzheimer’s biomarker discovery and protein structure prediction