LG Unveils ChatExaone AI Beta for Business

Published Date : 23-07-2025 

LG Group, South Korea’s fourth-largest conglomerate, has made its AI chatbot, ChatExaone, available to external users, marking its entry into the B2B AI segment. The tool, previously used internally, is now accessible to enterprise and institutional customers, rivaling leading models like OpenAI’s ChatGPT and Google’s Gemini 2.5. 

South Korea’s fourth-largest conglomerate, LG Group, has made its artificial intelligence chatbot, ChatExaone, available to external users, signaling its entry into the business-to-business (B2B) AI segment. The tool, previously deployed only for internal use to verify its accuracy and efficiency, is designed for enterprise and institutional customers rather than individual consumers – a strategic divergence from widely known services such as OpenAI’s ChatGPT.

LG AI Research Institute, LG Group’s AI think-tank, launched a public beta of the enterprise AI agent ChatExaone at LG Science Park in Seoul. This flagship event was meant to showcase the company’s AI ambitions. Companies, public agencies, and academic institutions can now access the platform, which requires verification via corporate email, underlining its business-oriented positioning, according to LG.

An LG executive said that ChatExaone’s performance is on par with leading global models like OpenAI’s GPT-o3 and Google’s Gemini 2.5 in terms of factual accuracy, logical reasoning, and answer consistency. The move is seen as a significant step by the conglomerate to commercialize its proprietary large language model (LLM), Exaone, and compete in the rapidly growing enterprise AI space.

We’ve been stress-testing Exaone within our affiliates since 2021. It is now mature enough to support external use across industries, said the LG executive. Since Exaone’s debut three years ago, LG has integrated the LLM across its subsidiaries, including LG Electronics Inc., LG Chem Ltd., and LG CNS Co., and has secured over 100 billion won ($72 million) worth of AI-related business contracts from its sister firms.

As generative AI continues to disrupt knowledge work and automation, LG’s strategic turn toward enterprise AI signals its ambition to play not just as a manufacturer but as a software-first innovator in the AI age. LG’s ambitious AI initiative was led by Group Chairman Koo Kwang-mo, who called for the creation of a “super-scale AI” capable of transforming operations across LG’s traditionally hardware-focused portfolio.

The result was the establishment of the LG AI Research Institute and the subsequent development of the Exaone LLM. At Tuesday’s media event, LG AI Research showcased its latest AI models, including Exaone 4.0 VL with multimodal capabilities that allow it to understand not just text but also images and video, enabling it to parse graphs and visuals embedded in documents. LG said Exaone 4.0 VL’s performance exceeds that of Meta’s LLaMA 4 Scout model, one of the benchmark offerings in enterprise-grade AI.

A specialized healthcare variant, Exaone Path 2.0, also presented at the event, uses medical imagery to deliver near-instant diagnostic support, capable of detecting conditions within a few minutes. At the event, LG executives emphasized its full-stack approach to AI, including Exaone On-Premise – a fully private deployment option that offers proprietary servers for enterprise customers concerned about data privacy.

The solution is powered by Korean AI semiconductor start-up FuriosaAI, with which LG is co-developing a neural processing unit (NPU) optimized for Exaone’s workloads. LG has also forged a partnership with FriendliAI, a Korean generative AI infrastructure company. Data generation is also part of the broader offering. With its Exaone Data Foundry, LG said it can reduce data creation cycles dramatically.

What used to take a team of 60 people three months can now be done by one person in a single day, said another LG executive, citing pilot projects with government and industry clients that yielded a 1000-fold increase in productivity and 20% improvement in data quality. One of the most tangible outcomes of LG’s AI push is a joint project with the London Stock Exchange Group (LSEG), announced at the event.

Starting in the third quarter, the two companies agreed to launch AI-generated investment reports for less-covered stocks such as small-caps and nanocaps – a use case that could support M&A activities and investor research projects. Analysts said the partnership with LSEG marks a significant leap for LG into high-value AI services with global implications. We are building an industrial-grade AI ecosystem, said Lee Hong-lak, chief AI scientist at LG and co-head of LG AI Research Institute. Our goal is to deliver safe, high-performance models that can be deeply embedded into real-world operations. 

Frequently Asked Questions (FAQS): 

Q: What is ChatExaone?

A: ChatExaone is an AI chatbot developed by LG Group. It is designed for enterprise and institutional customers, offering high performance in terms of factual accuracy, logical reasoning, and answer consistency.

Q: How does ChatExaone compare to other AI models?

A: ChatExaone’s performance is on par with leading global models like OpenAI’s GPT-o3 and Google’s Gemini 2.5, according to LG executives.

Q: What is Exaone?

A: Exaone is LG’s proprietary large language model (LLM) that has been integrated across its subsidiaries. It is now available for external use in the B2B AI segment.

Q: What are the key features of Exaone 4.0 VL?

A: Exaone 4.0 VL has multimodal capabilities, allowing it to understand text, images, and video. It can parse graphs and visuals embedded in documents and outperforms Meta’s LLaMA 4 Scout model.

Q: What is the partnership between LG and the London Stock Exchange Group (LSEG)?

A: LG and LSEG have agreed to launch AI-generated investment reports for less-covered stocks such as small-caps and nanocaps, which could support M&A activities and investor research projects. 

More Related Topics :