An infographic shows the timeline and second-round evaluation results of South Korea’s “Independent AI Foundation Model” project. Upstage, SK Telecom and LG AI Research advanced to the next round, while Motif Technologies was eliminated. The government plans to select the final two teams in early 2027. Data from Ministry of Science and ICT. Graphic by Asia Today and translated by UPI
Aug. 18 (Asia Today) — South Korea’s competition to develop an independent national artificial intelligence foundation model has narrowed to three teams, with LG AI Research, SK Telecom and Upstage advancing to the next round, the government said Tuesday.
Motif Technologies, which joined the project through an additional competition in February, ranked last in the second-stage evaluation and was eliminated.
The Ministry of Science and ICT launched the Independent AI Foundation Model project last September with five teams as part of an effort to reduce dependence on foreign AI models.
The government initially set a target of developing a Korean foundation model reaching about 95% of the performance of frontier-class AI models.
Naver Cloud and NC AI were eliminated in the first evaluation in January. An additional application process in February brought Motif Technologies into the contest, leaving four teams competing in the second round.
The second-stage evaluation assigned 40 points to benchmark performance, 35 points to expert evaluations and 25 points to user evaluations.
The ministry said the process was more sophisticated than the first evaluation, incorporating difficult international benchmarks, public participation in usability testing and expanded assessment of how the models could be deployed in industrial AI transformation.
The four teams averaged 22.5 points in benchmark testing, with a 4-point difference between first and fourth place.
They averaged 28.8 points in expert reviews, with a 2.4-point gap between first and fourth place.
The average user-evaluation score was 17.6 points and the gap between first and fourth place was 5 points.
All four models were also determined to have achieved technological independence, according to the ministry.
Second Vice Minister of Science and ICT Ryu Je Myung said different teams placed first in each evaluation category.
“In the case of Motif Technologies, it demonstrated a world-class level from a technological standpoint, but it received lower evaluations than the other models in usability and practical application,” Ryu said.
The government’s official English-language materials identify him as Ryu Je Myung, second vice minister of science and ICT.
The three remaining teams will continue developing their AI models before a third evaluation scheduled for early next year, when two teams are expected to be selected.
The ministry said it plans to increase support as global competition in frontier AI intensifies.
It plans to expand the number of Nvidia B200 high-performance graphics processing units provided to the teams from about 768 during the first half of this year to roughly 1,000.
The government is also reviewing a new support system and plans to announce additional details later.
Ryu said the ministry and participating companies have also discussed whether the current structure of the project is sufficient as global frontier AI companies rapidly improve their models.
“We have continued discussions with the companies about whether the current scale of support and competitive structure of the project can keep pace as the performance of global frontier companies’ models improves exponentially,” Ryu said.
“There is a shared understanding that we should move away from the current approach and reorganize the program in a way that allows competition at a level comparable to frontier companies’ models.”
— Reported by Asia Today; translated by UPI
© Asia Today. Unauthorized reproduction or redistribution prohibited.
Original Korean report: https://www.asiatoday.co.kr/kn/view.php?key=20260818010005838