SEONGNAM, SOUTH KOREA -- 2Digit, a Korean AI startup led by CEO Edward Park, announced today that it has become the first company globally to surpass the 80-point barrier on the Open LLM Leaderboard, operated by the open-source platform Hugging Face. It has maintained its first-place position for 57 consecutive days.

The Open LLM Leaderboard is a prestigious international AI competition hosted by Hugging Face, the world's largest open-source platform. Companies participate by selecting their AI language models and are scored based on their fine-tuning capabilities, with rankings determined accordingly.

2Digit specializes in providing news recommendation services based on natural language processing (NLP) AI tailored for news content. Their core service involves highly personalized recommendations, classifying and suggesting large volumes of news data according to user interests.

In this competition, 2Digit achieved a significant milestone by breaking the 80-point barrier, a feat that neither Korean nor global competitors had previously managed. As of May 20, it holds the top position, followed by Russia's MTS AI Lab in second and third places, and San Francisco-based ABACUS.AI in fourth. Other Korean companies include MORE in 10th place and Saltlux in 16th, with Upstage positioned outside the top 150.

CEO Edward Park of 2Digit commented, “The claim that AI companies can create foundation models and use them universally as a panacea is unrealistic. The true measure of an AI company's technology lies in its ability to optimize and fine-tune models for specific business needs. It is regrettable that the importance of foundation models is sometimes overstated, overshadowing the necessity and value of fine-tuning technologies.”

Additionally, 2Digit secured first place in the Korean LLM competition organized by NIA (National Information Society Agency) last March and was selected for KT‘s ’Large-scale AI Utilization Support Project,' where they developed an economic news reading service, which was recognized as an exemplary case by KT.