Friendly AI-LG Debuts 'Exaone 4.0' Globally

LG AI Research's latest language model, 'EXAONE 4.0,' is making its global debut through FriendlyAI's serverless inference platform, boasting a cost efficiency at one-tenth the level of ChatGPT. EXAONE 4.0, specialized in professional domains, was developed to be optimized for various enterprise environments, based on its excellent reasoning capabilities and natural language processing performance. Companies can easily integrate EXAONE 4.0 into their services in the form of an API provided by FriendlyAI, without needing to build separate GPU infrastructure. FriendlyAI's inference platform provides fast response times and automatic scaling capabilities based on traffic, supporting companies in reliably adopting EXAONE 4.0 for large-scale production services. Lee Hwa-young, Managing Director at LG AI Research, stated that as EXAONE 4.0 has achieved a global standard in general and inference domains, expectations for its utilization in industrial sectors are high. She emphasized that FriendlyAI's flexible and high-performance inference infrastructure will be a crucial turning point for the domestic and international expansion of EXAONE 4.0. Jeon Byung-gon, CEO of FriendlyAI, said that through this collaboration, the powerful performance of EXAONE can be utilized at approximately one-tenth the cost of ChatGPT, which will significantly improve accessibility to AI technology. FriendlyAI, which demonstrated its technological prowess by signing an official model distribution partnership with Hugging Face, the world's largest AI model hub, earlier this year, is expected to create synergy with LG AI Research's model development capabilities, breaking down technical and cost barriers to AI adoption and playing a key role in accelerating its widespread dissemination.