Unicorn company Lingyiwanwu in the field of large-scale models in China has officially launched the Yi Big Model API Open Platform today, providing developers with more choices and possibilities. The Yi Big Model API Open Platform brings together a variety of high-performance models, such as Yi-34B-Chat-0205, which supports various functions such as general chat, question answering, dialogue, writing, and translation. Yi-34B-Chat-200K has the ability to process 200K contexts and performs well in areas such as multi-document reading comprehension and building super-long knowledge bases. Yi-VL-Plus is a multimodal model that supports text and visual multimodal inputs, and its Chinese chart experience even surpasses GPT-4V. It is worth mentioning that the Yi Big Model API Open Platform has started invitation testing half a month ago, and many developers have already experienced the powerful capabilities of these models. Among them, Yi-34B-Chat-200K performs particularly well in translating professional books. A well-known technical author completed a translation of a 264-page book in just one day. In addition, the Yi Big Model API Open Platform is also compatible with the emerging giant OpenAI API, providing great convenience for migration solutions. However, whether the Yi Big Model API can compete with top models such as GPT-4 Turbo, Gemini 1.5, and Claude 3 still needs further verification from more developers. The launch of the Yi Big Model API Open Platform by Lingyiwanwu undoubtedly brings a high-performance multimodal large-scale model to the Chinese community. Especially Yi-VL-Plus, it demonstrates powerful potential in productivity scenarios and can become an effective assistant for user analysis of charts, knowledge classification, and data summarization, significantly improving work efficiency. Today, the victory or defeat of large-scale model manufacturers in competition no longer solely depends on cold ranking data. Reducing the threshold for model usage and reducing the burden on users are equally important. Therefore, open APIs have become a common choice for many manufacturers. Since its establishment, Lingyiwanwu has insisted on open sourcing the Yi series models to the public and contributing to the open-source community. At the same time, through open APIs, more developers can easily utilize powerful conversational and multimodal large-scale models, thereby promoting the application of these models in more scenarios. Looking ahead, Dr. Kai-Fu Lee, CEO of Lingyiwanwu, has stated that the company will create more consumer-oriented super applications based on the Yi series of large-scale models. While opening up dialogue and multimodal model APIs, Lingyiwanwu also emphasizes the important role of developer tools in promoting innovation in large-scale model applications. With these dual efforts, Lingyiwanwu has laid a solid foundation for achieving its grand goals. It is reported that Lingyiwanwu will also provide developers with more powerful models and AI development frameworks in the near future, covering larger parameter sizes, stronger multimodal capabilities, and more professional code/mathematical reasoning models. At the same time, the company will strive to break through longer context processing capabilities, improve inference speed, and reduce inference costs. Based on these technological breakthroughs, Lingyiwanwu will build a new developer AI framework, including vector databases, RAG, and Agent architecture, to provide developers with richer and more flexible tools to meet the needs of diverse application scenarios. It is evident that Lingyiwanwu has formed a clear strategic thinking in the development of large-scale models and will continue to delve deeper and go further in the future. We look forward to Lingyiwanwu bringing more innovative technological achievements and driving continuous development in the field of large-scale models.