LG AI Research

🚀LG AI Research and NVIDIA are deepening their strategic alliance to expand the 'EXAONE' ecosystem by co-developing domain-specific models integrated with NVIDIA Nemotron.

💡 Building on a successful history from EXAONE 3.0 to 4.5 and K-EXAONE this partnership leverages NVIDIA’s Blackwell GPUs and NeMo framework to deliver world-class AI performance.

🌍 Together, we are accelerating 'Sovereign AI' initiatives to drive real-world industrial transformation and solidify South Korea’s leadership in the global AI landscape.

At American Association for Cancer Research 2026 ( ), LG AI Research and Vanderbilt University Medical Center ( ) presented a set of researches at the intersection of spatial biology and AI—spanning 3D spatial multi-omics, real-time 4D drug response tracking, ROI optimization, and foundation models linking pathology with molecular data.

Rather than focusing on single modalities or static snapshots, this work explores how to reconstruct tumor microenvironments in their true context, and how to translate complex spatial data into actionable biological insight.

🔗Explore more on our blog:
[ENG] https://bit.ly/4tqlMyx
[KOR] https://bit.ly/4vI5P8q

Big ideas in cancer research are heading to San Diego 🚀 As the global oncology community gathers at American Association for Cancer Research ( ) Annual Meeting 2026 from April 17–22, LG AI Research is bringing the latest research from our Bio Intelligence Lab.

Covering gene expression foundation models, cancer-scale foundation model, and omics-aware patch aggregation, we’ll be sharing how AI is unlocking deeper insights into cancer biology.

Come visit us at Booth #4600 and meet the team!
🎤Presenters: Minsoo Lee, Jong Hyun Kim, Jongseong Jang
👉 Discover more: https://bit.ly/3QdoN6u

What happens when understands both text and visuals as one? 4.5, the latest model from LG AI Research, expands beyond language by integrating visual and textual understanding within a single architecture.

Built with native multimodal pretraining, it processes images and text together from the ground up, enabling more contextual and accurate reasoning across real-world scenarios.

This marks a step toward more integrated AI.

Explore the next evolution of EXAONE on our newsroom!
(ENG) https://bit.ly/4tDqVTu
(KOR) https://bit.ly/48LaEUr

Taking on the Challenge with ‘EXAONE,’ Korea’s Representative AI
LG Aimers 8th participants dived deep into the world of AI Model Compression using Korea’s representative AI, ‘EXAONE.’ From coding through the night to 1:1 career consulting with LG experts, it was a journey of pure innovation and growth.

Congratulations to everyone who pushed their limits!

🔗Read the full story on our blog:
[ENG] https://bit.ly/3PS9U9F
[KOR] https://bit.ly/48cgU7u

🔗Check out the highlights from the event on YouTube: https://bit.ly/4scps5p

📍Next up: LG Aimers 9th recruitment starts this May : https://bit.ly/47FjOBx

#해커톤

👀 As generates and interprets images at scale, what matters most is distinguishing what’s real—and reasoning from what’s seen.

From detecting AI-generated images with to mitigating bias in VLMs through , these innovations address key challenges in AI.

Featured in our LG Accountability Report on AI Ethics 2025, these efforts reflect our commitment to building more trustworthy AI systems.

What do you think about these approaches? Share your thoughts in the comments below 👇

Want to learn more? Check out the full report and our blog.
🔗Full report: https://bit.ly/4sLAfES
🔗[ENG] https://bit.ly/41E7YUE
🔗[KOR] https://bit.ly/4tDlgNt

The era of simple adoption is over. 🏭The future of the Autonomous Economy demands transparency and trust.

At AAAI 2026, the answer was clear: Move beyond black-box models to a Transparent Architecture built on Ontology.

Now is the golden time to establish a Trustworthy Agent Strategy. LG AI Research is leading the way, engineering AI that isn't just powerful—but explainable and efficient.

Check out our full insights from AAAI 2026 on the blog! 🔗
[ENG] https://bit.ly/4bLiDBM
[KOR] https://bit.ly/4bE7Lpa

Check out LG AI Research’s AAAI 2026 paper! 🔗 https://bit.ly/4d0LyUG