"'K-Exaone,' which is participating in the government's national AI foundation model project, is delivering performance on par with OpenAI. We aim for a global best-in-class level." (Im Woo-hyeong, co-head of research at LG AI Research)
"Even for a Humanoid Robot ultimately built on the AI agent 'ixi-O,' the way it communicates will be voice. We are preparing to apply diverse voice technologies to physical AI going forward." (Lee Sang-yeop, chief technology development officer at LG Uplus)
LG Uplus Chief Technology Development Officer (CTO) Lee Sang-yeop and LG AI Research co-head of research Im Woo-hyeong held a briefing on the 1st, a day before the opening of the world's largest mobile communications expo, 'MWC 26,' in Barcelona, Spain, and stated accordingly the blueprint for 'No. 1 AI One Team LG.' LG Uplus's ixi-O is equipped with Exaone.
K-Exaone, developed in-house by LG AI Research, recently took first place in phase one of the government's national AI project screening. However, it is not No. 1 on the global stage yet. According to the global AI model performance analysis organization Artificial Analysis, as of last January, based on open-weight models that disclose weights, K-Exaone scored 32 points, ranking seventh in the world. OpenAI ranked sixth with 33 points. LG plans to develop the highest-performing language model among existing global open-weight models during the second phase of its independent AI foundation model project in the first half of this year.
On this day, LG Group unveiled for the first time the future development roadmap of 'K-Exaone (EXAONE),' which is being developed as a national AI model. LG plans to pursue Exaone development around four key strategies: securing leadership in AI foundation model development; aiming for expert AI; expanding practical, on-site industrial applications; and securing trust and safety for sustainable AI.
Im Woo-hyeong, co-head of research at LG AI Research, said, "The AI LG pursues is not a competition for how high the intelligence is, but about creating a partner that helps human life and solves problems in the real world," adding, "It is important to build good AI, but what we ultimately build is the core. We will go beyond the stage of AX (AI transformation) and focus on creating AI that generates real value in the physical space of the real world."
They also outlined plans to connect K-Exaone with infrastructure. LG Uplus CTO Lee Sang-yeop said, "To implement AI that can help in the real world, powerful infrastructure and connection technologies that support it are essential," adding, "As K-Exaone moves into its completion phase next year, One Team LG, including LG Uplus, is preparing Korea's largest AI data center (DC)." He continued, "We are proceeding without a hitch so we can deliver the best experience with the world's highest-performing AI," introducing the Paju AI DC, the largest in the capital region (200 MW), slated for completion in 2027.
The Paju AI DC can house up to 120,000 GPUs. The site is expected to become the execution ground for the One LG strategy, which consolidates the core capabilities of LG affiliates such as LG Electronics, LG Energy Solution and LG CNS. LG Uplus plans to provide complete end-to-end, customized services for customers based on 'K-Exaone' and the 'Paju AIDC.'
In addition, LG plans to showcase infrastructure that combines Exaone from LG AI Research, LG Uplus's enterprise AI platform, and FuriosaAI's Neural Processing Unit (NPU) to demonstrate the synergy between K-AI and K-semiconductors.
LG AI Research plans to soon release 'Exaone 4.5,' a vision-language model (VLM), as an open-weight model as it nears completion. A vision-language model is multimodal technology that combines language intelligence and visual intelligence to understand, analyze and communicate with text and visual information in an integrated, humanlike way. The company said it expects 'Exaone 4.5' to increase the model's utility through modality expansion while becoming the highest-performing model globally among open-weight models of the same size. In particular, starting with 'Exaone 4.5,' LG expects that the Exaone vision-language model technology to be advanced going forward will serve as the brain for the Korea-style humanoid 'KAPEX,' taking root as a core technology that will usher in the era of physical AI.
On this day, LG Uplus also introduced a next-generation agentic AI strategy it will develop in partnership with LG AI Research. LG Uplus CTO Lee Sang-yeop said, "As a telecom operator, we aim to evolve beyond simply connecting networks to becoming an AI software assistant."
In addition, LG Uplus will build a Hybrid AI Infra architecture optimized to run agentic AI in real time. The architecture is expected to become a core technology that enables stable and cost-effective operations with minimal latency for large language models such as Exaone within corporate internal environments.
LG Uplus CTO Lee Sang-yeop said, "Real competitiveness in the AI era does not lie merely in a volume offensive to secure bigger models and more compute resources," adding, "The key battleground will be how precisely we design an agentic architecture that evolves on its own through a cyclical loop of going beyond single inference to repeated planning and execution, evaluation and correction."