The 2023 Global Large Language Model Competition has begun, with a surge of Chinese large language models. The Chinese evaluation platform CLEVA has been launched, featuring comprehensive evaluation tasks and metrics. CLEVA focuses on indicators such as accuracy, robustness, fairness, efficiency, calibration, and diversity. A variety of prompt templates ensure fairness in evaluation and analysis of model performance. CLEVA employs multiple methods to mitigate data contamination risks, providing a clear interface for user interaction.