Krafton has had 10 papers accepted at one of the world's leading artificial intelligence (AI) conferences, highlighting the recognition of its research in multimodal large language models (LLMs) and optimization techniques.
On July 11, Krafton announced that it presented a total of 20 papers, including 10 in the main track and 10 in the workshop track, at the International Conference on Machine Learning (ICML) 2026, held at COEX in Seoul from July 6. ICML is considered one of the top three AI conferences globally, and Krafton has achieved its highest performance at these three major AI events.
The accepted papers cover key research findings related to Krafton's foundation model, including world models, multimodal LLMs, preference learning, reasoning, and optimization.
Krafton AI has been actively presenting its deep learning research papers at global AI conferences. In total, 85 papers from Krafton AI have been accepted at the three major conferences: NeurIPS, ICML, and ICLR.
In December 2023, Krafton presented five main track papers at NeurIPS 2023, the highest number among domestic gaming companies and the third highest overall among international gaming firms.
In 2024, Krafton presented a total of eight main track papers at ICLR, ICML, and NeurIPS. In 2025, the company presented 15 main track papers, and this year, it has published a total of 17 papers.
Additionally, during ICML, Krafton hosted an AI networking event called 'AI for Games' to promote collaboration within the ecosystem at the intersection of gaming and AI.
A Krafton representative stated, “We plan to take the lead in securing technological competitiveness that can contribute to actual games and serve as a catalyst for the expansion of the domestic gaming industry and AI ecosystem.”
Meanwhile, Krafton is nurturing AI as a core growth driver for the next generation. The company has established a research and development system that spans from basic deep learning research to application in game services. It is pursuing a strategy to utilize AI research outcomes in game production, AI characters, and multimodal technologies.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.