ChatLaw is an LLM-based multi-agent legal assistant enhanced by knowledge graphs and Mixture-of-Experts architecture. It offers specialized models including ChatLaw2-MoE (InternLM-based 4x7B), ChatLaw-13B (Ziya-LLaMA-based), and ChatLaw-33B (Anima-based), all fine-tuned for Chinese legal language processing. The project also includes ChatLaw-Text2Vec, a text similarity model trained on 93,000 court cases for matching user queries to relevant legal statutes.