 | | 0 |
| Source: Yonhap, Ministry of Science and ICT/ Graphic by AsiaToday's Park Jong-gyu |
South Korea’s first-round evaluation of an “independent AI foundation model” project has narrowed the field to three teams, with LG AI Research earning the highest score, while Naver Cloud and NC AI failed to advance.
The Ministry of Science and ICT, together with the National IT Industry Promotion Agency and the Telecommunications Technology Association, said Thursday that LG AI Research, SK Telecom, and Upstage will move on to the second stage of the government-backed initiative aimed at building sovereign AI capabilities.
In the comprehensive assessment, which combined benchmark testing, expert review, and user evaluation, LG AI Research topped the list with a total score of 90.2 out of 100. The average score among the five elite teams was 79.7.
Benchmark results showed LG AI Research and SK Telecom tied for the highest domestic score, while LG led global benchmark categories. LG also ranked first in expert reviews—conducted by a 10-member panel from academia, industry, and research institutes—and achieved a perfect score in user evaluations.
Vice Minister Ryu Je-myung said the assessment examined not only model performance and real-world usability, but also cost efficiency, ecosystem impact, and independence, including technical, policy, and ethical considerations.
Naver Cloud was ruled out after failing to meet independence criteria in technical and policy aspects, according to the ministry, echoing concerns raised by expert reviewers. As a result, only LG AI Research, Upstage, and SK Telecom will proceed to the next phase of development.