Korea AI Scores: Motif Controversy & Evaluation Details

South Korea's Ministry of Science and ICT has released more detailed scores from its "Sovereign AI Foundation Model" project. This follows ongoing controversy after Motif Technologies, which had the highest global AI benchmark score, was eliminated. Here's the breakdown: the second-round evaluation used a 100-point scale. This included 40 points for benchmarks, 35 for expert evaluation, and 25 for user evaluation. Motif Technologies ranked first in the global AI performance index with 11.9 points out of 25, while the four-team average was 9.48. Motif 3 had previously scored 47 on the raw index, ahead of other contenders. However, SK Telecom received the highest score in the domestic benchmark, with 13.4 points out of 15. LG AI Research led the expert evaluation with 29.5 points out of 35. SK Telecom also ranked first in the professional user assessment, scoring 11.6 points. LG AI Research topped the general public assessment with 7.6 points out of 10. In summary, Motif led the global benchmark, SK Telecom ranked first in domestic and professional user evaluations, and LG AI Research topped the expert and general public evaluations. Upstage did not rank first in any of these categories. The bottom line is this detailed disclosure aims to address transparency concerns surrounding the project's evaluation process.