LG Launches South Korea’s First Open-Source AI Model Exaone 3.0

South Korea has entered the global AI landscape with the launch of LG AI Research's Exaone 3.0, the country’s first open-source AI model. The 7.8 billion parameter model excels in both Korean and English language tasks and aims to accelerate AI research and contribute to building a robust AI ecosystem in Korea.

As stated in arxix.org, “EXAONE stands for EXpert AI for EveryONE, a vision that LG is committed to realizing in order to democratize access to expert-level artificial intelligence capabilities.”

The research paper also states that the objective of Expert AI is twofold: to help the general public achieve expert-level competency in various fields and to assist experts in attaining even higher levels of proficiency.

Notably, Exaone 3.0 promises improved efficiency, as LG claims a 56% reduction in inference time, a 35% decrease in memory usage, and a 72% reduction in operational costs compared to its predecessor.

Further, the model has been trained on 60 million cases of professional data related to patents, codes, math, and chemistry. The company plans to expand model training to 100 million cases across various fields by year-end.

With this move, LG positions itself at the forefront of AI innovation. Through Exaone 3.0, LG could anticipate a new revenue stream in cloud computing and AI services.