AGI-Eval is a professional community focused on evaluating the capabilities of large AI models, aiming to provide users with model performance references through systematic evaluations.

LMArena is a crowdsourced comparison-based AI model evaluation platform that measures the actual performance of large language models through real-world blind dialogue tests.

The open-source large model performance benchmark leaderboard maintained by Hugging Face provides a transparent and standardized quantitative comparison of model capabilities.