Google Cloud Supports Development of LG AI Model 'Exaone 3.0'

Google Cloud announced on the 28th that LG AI Research has built its latest version of the generative AI model, 'EXAONE 3.0,' and the enterprise AI agent 'ChatEXAONE' based on Google Cloud's AI-optimized infrastructure and solutions.

LG AI Research has maintained close collaboration with Google Cloud since the early development of EXAONE. In 2021, to develop and enhance the training of the EXAONE 1.0 model, they introduced Google Cloud TPU (AI specialized chips) and established an efficient model training process.

Additionally, during the development of EXAONE 3.0, they further enhanced compatibility for training and inference by incorporating Google Cloud's Graphics Processing Units (GPUs).

They plan to expand business cooperation by offering EXAONE 3.0 on the Google Cloud AI Marketplace and introducing LG AI Research's AI models to other Google Cloud customers.

Lee Hwa-young, Executive Director of LG AI Research, stated, "Thanks to Google Cloud's world-class AI infrastructure and its flexible, cost-effective architecture that can be customized to meet specific demands, we were able to successfully launch EXAONE 3.0."

Jigi-seong, President of Google Cloud Korea, said, "Google Cloud has maintained close collaboration with LG AI Research for years, from the initial development and enhancement stages of EXAONE 1.0. We look forward to continuing cooperation with LG AI Research, which is leading Korea's core AI technology, across various fields and opening new horizons for generative AI."