About LiveBench
LiveBench is a benchmark suite for large language models (LLMs) that prioritizes objective and contamination-free evaluation. Designed for researchers, developers, and organizations, LiveBench regularly releases new test questions to minimize test set leakage and ensures all benchmark tasks have objectively verifiable answers. The benchmark covers diverse tasks—such as reasoning, coding, mathematics, data analysis, and language—and maintains an up-to-date leaderboard showcasing LLM performance across leading AI providers.
