Llama3.1-8B-Chinese-Chat is an instruction-tuned language model based on the Meta-Llama-3.1-8B-Instruct architecture, designed for both Chinese and English users. It offers multiple capabilities such as role-playing and tool usage. The model is fine-tuned using the ORPO algorithm, significantly reducing the incidence of Chinese questions answered in English and minimizing language mixing, particularly excelling in role-playing, functionality invocation, and mathematical reasoning.