Introducing the LiveCodeBench Leaderboard - Holistic and Contamination-Free Evaluation of Code LLMs

LiveCodeBench introduces a new leaderboard for evaluating code-focused Large Language Models (LLMs) with an emphasis on holistic assessment and contamination-free testing. The benchmark aims to provide more accurate and reliable evaluation of AI coding capabilities by addressing common issues in existing evaluation methods.