This survey provides a comprehensive overview of recent advances and challenges in multilingual large language models (LLMs). It discusses the development of LLMs, their training methods, and the challenges in multilingual scenarios, including data scarcity, knowledge transfer, and domain adaptation. The survey highlights the importance of multilingual capabilities in LLMs for improving accessibility and usability across diverse language communities. It also explores various inference strategies, security concerns, and future research directions for enhancing multilingual LLMs. The survey emphasizes the need for systematic categorization and comparison of existing LLMs to address the challenges in multilingual scenarios. It also discusses the limitations of current LLMs and proposes future research directions to improve their multilingual capabilities. The survey concludes with a call for further research to develop more effective and efficient multilingual LLMs.This survey provides a comprehensive overview of recent advances and challenges in multilingual large language models (LLMs). It discusses the development of LLMs, their training methods, and the challenges in multilingual scenarios, including data scarcity, knowledge transfer, and domain adaptation. The survey highlights the importance of multilingual capabilities in LLMs for improving accessibility and usability across diverse language communities. It also explores various inference strategies, security concerns, and future research directions for enhancing multilingual LLMs. The survey emphasizes the need for systematic categorization and comparison of existing LLMs to address the challenges in multilingual scenarios. It also discusses the limitations of current LLMs and proposes future research directions to improve their multilingual capabilities. The survey concludes with a call for further research to develop more effective and efficient multilingual LLMs.