Upstage Tops HuggingFace Open LLM Leaderboard, Outscoring ChatGPT in Language Model Evaluation

South Korean AI startup Upstage has achieved a milestone by surpassing OpenAI’s ChatGPT, one of the most well-known chatbot models, in a widely recognized open-source language model evaluation. Upstage’s new AI model, trained with Meta’s LLaMA 2 70B model, has obtained an average score of 72.3 on the latest HuggingFace Open LLM Leaderboard. This marks the first time a non-major AI model has outperformed ChatGPT, which has a benchmark score of 71.9. The victory demonstrates Upstage’s technological competence on the global AI stage.

Previously, Upstage’s LLM model with 30 billion…