Kakao Releases 4 Lightweight 'Kanana-2' AI Models as Open Source, Boosting Korean Language Processing Efficiency by 30%
Kakao has officially released four lightweight artificial intelligence (AI) models, collectively named ‘Kanana-2’, designed to run efficiently on personal devices such as smartphones and PCs. This release is seen as a strategic move to accelerate the popularization of high-performance AI technology and support developers in building optimized AI services across various environments.
According to industry reports, the core feature of the newly introduced Kanana-2 models is a 30% improvement in Korean language processing efficiency compared to previous versions. Specifically, the design is optimized for on-device AI environments, allowing for fast response times and high accuracy directly on the device without the need for cloud server processing.
The four models released by Kakao offer varying levels of computational power and memory usage, allowing users to flexibly choose based on their device specifications. This reflects Kakao’s technical commitment to overcoming the limitations of Large Language Models (LLMs) that typically require high-end servers, enabling high-quality AI experiences on consumer mobile devices.
This open-source release is expected to have a significant impact on the domestic and international AI ecosystem. As a model specialized for the Korean language, it is expected to be a valuable asset not only for domestic developers but also for global companies preparing Korean-based AI services. Kakao plans to lower the barrier to entry for AI technology and strengthen cooperation with the open-source community through this release.
Technically, despite being lightweight, Kanana-2 maintains the ability to understand complex contexts and generate natural language. This is the result of combining Kakao’s accumulated know-how in large-scale language data processing with efficient model compression technology. Its ability to precisely handle the complex grammar and nuances of the Korean language is a key differentiator from other lightweight models competing in the global market.
The open-sourcing of Kanana-2 is considered an important milestone in Kakao’s goal of ‘AI popularization.’ Beyond simply possessing technology, it reflects an intention to enhance the competitiveness of the Korean AI industry by raising the overall technical level of the ecosystem. Furthermore, as developers are now free to modify and distribute the models, the emergence of various derivative services and innovative applications is expected.
In conclusion, Kakao’s move is a strategic step to set the standard for Korean language processing technology in the global AI market and lead the era of on-device AI. At a time when AI efficiency and data privacy are becoming globally critical, the emergence of high-efficiency models that run on-device will serve as an opportunity to provide global users with new digital experiences that combine K-content and AI technology.
