Kakao is set to release Kanana 2.5, a large language model (LLM) with 150 billion parameters (150B).


Kakao Pangyo Ajit in Seongnam, Gyeonggi Province. Photo by Jinhyung Kang aymsdream@

Kakao Pangyo Ajit in Seongnam, Gyeonggi Province. Photo by Jinhyung Kang aymsdream@

View original image

Shina Jung, CEO of Kakao, stated during the company's first-quarter earnings conference call held on the morning of May 7, "We are about to release the Kanana 2.5 model," adding, "Kakao developed this model from scratch using its own technology, and we have confirmed that it demonstrates the best performance among domestic and international LLMs of a similar scale."



Jung further emphasized, "Last year, we developed our own tokenizer, which minimized training costs and maximized inference speed," and stated, "Kanana 2.5 possesses the most efficient Korean language compression capability among all domestic and international models released to date."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.

Today’s Briefing