Introducing the Open Ko-LLM Leaderboard: Leading the Korean LLM Evaluation Ecosystem
Hugging Face
Upstage launched the Open Ko-LLM Leaderboard in September 2023 to evaluate Korean-language large language models through standardized benchmarks and a public submission platform. The leaderboard has received over 1,000 model submissions in its first five months, compared to 4,000 on the English-language Open LLM Leaderboard, with participants including major corporations like KT and academic institutions such as KAIST and Korea University. The platform enables researchers to benchmark Korean LLMs against five evaluation tasks and identify top-performing models, while revealing infrastructure limitations that cause delays for models exceeding 30 billion parameters.