This paper provides a comprehensive review of the evolution from Large Language Models (LLMs) to Large Multimodal Models (LMMs), highlighting their development, key techniques, and applications. It begins by summarizing the progress of LLMs, including their architectures, pretraining, fine-tuning, and prompt engineering. The paper then focuses on LMMs, discussing their architectural components, training strategies, instruction tuning, and prompt engineering. It presents a taxonomy of the latest 66 vision-language LMMs, analyzing their different types. Finally, the paper offers a unified perspective on both LLMs and LMMs, analyzing the development status of large-scale models globally and offering potential research directions. The review emphasizes the similarities and differences between LLMs and LMMs, and highlights the importance of a unified perspective in understanding their evolution. It also addresses the global development of large models, noting disparities in development across regions and the need for a comparative analysis. The paper concludes by summarizing the key findings and future directions for large-scale models.This paper provides a comprehensive review of the evolution from Large Language Models (LLMs) to Large Multimodal Models (LMMs), highlighting their development, key techniques, and applications. It begins by summarizing the progress of LLMs, including their architectures, pretraining, fine-tuning, and prompt engineering. The paper then focuses on LMMs, discussing their architectural components, training strategies, instruction tuning, and prompt engineering. It presents a taxonomy of the latest 66 vision-language LMMs, analyzing their different types. Finally, the paper offers a unified perspective on both LLMs and LMMs, analyzing the development status of large-scale models globally and offering potential research directions. The review emphasizes the similarities and differences between LLMs and LMMs, and highlights the importance of a unified perspective in understanding their evolution. It also addresses the global development of large models, noting disparities in development across regions and the need for a comparative analysis. The paper concludes by summarizing the key findings and future directions for large-scale models.