**国内大模型独角兽公司零一万物今日宣布，其Yi大模型API开放平台已正式上线，标志着人工智能技术在中文处理领域的又一重大突破。该平台首批推出了三个高性能模型，旨在为开发者和企业提供强大的文本处理能力，支持高达30万汉字的处理量。**
零一万物，由知名创新工场创始人李开复博士创办，一直致力于人工智能领域的前沿探索。此次发布的API开放平台，是公司技术实力的集中体现。平台提供的三个模型包括：
1. **Yi-34B-Chat-0205**：这是一个全能型模型，能够胜任通用聊天、问答、对话、写作及翻译等多种任务，为用户提供了全方位的自然语言处理解决方案。
2. **Yi-34B-Chat-200K**：该模型以200K的上下文能力脱颖而出，擅长多文档阅读理解，对于构建超长知识库的应用场景表现出色，是科研和企业级应用的理想选择。
3. **Yi-VL-Plus**：作为一款多模态模型，Yi-VL-Plus不仅支持文本输入，还能处理视觉信息，其在中文图表理解方面的性能超越了GPT-4V，为多媒体内容理解和生成提供了新可能。
零一万物Yi大模型API开放平台的推出，无疑将进一步推动中文智能应用的发展，为开发者带来更高效、更精准的工具，同时也为各行各业的数字化转型注入新的活力。欲了解更多详情，可访问平台官网：[https://platform.lingyiwanwu.com/](https://platform.lingyiwanwu.com/)。
英语如下：
**News Title:** “Innovator Li Kaifu’s Zero One Everything Launches Yi Large Model API: Multimodal Processing of 300,000 Chinese Characters, Pioneering a New Era in AI Interaction”
**Keywords:** Zero One Everything, Yi Large Model, API Open
**News Content:** **Zero One Everything, a domestic unicorn in large model development, announced today that its Yi Large Model API Open Platform is now live, marking another significant breakthrough in artificial intelligence technology for Chinese language processing. The platform initially offers three high-performance models designed to provide developers and enterprises with powerful text processing capabilities, supporting up to 300,000 Chinese characters.**
Founded by renowned Dr. Li Kaifu, the founder of Innovation Works, Zero One Everything has consistently focused on cutting-edge explorations in the field of artificial intelligence. This newly released API platform is a testament to the company’s technical prowess. The platform consists of the following three models:
1. **Yi-34B-Chat-0205:** This all-purpose model excels in general chatting, question-answering, dialogue, writing, and translation tasks, providing a comprehensive natural language processing solution for users.
2. **Yi-34B-Chat-200K:** With its standout 200K context capacity, this model is adept at multi-document reading comprehension, making it ideal for applications requiring extensive knowledge bases and for both research and enterprise-level use.
3. **Yi-VL-Plus:** As a multimodal model, Yi-VL-Plus not only handles text input but also processes visual information. Its performance in understanding Chinese charts surpasses that of GPT-4V, opening up new possibilities for multimedia content understanding and generation.
The launch of Zero One Everything’s Yi Large Model API Open Platform undoubtedly accelerates the development of Chinese intelligent applications, offering developers more efficient and precise tools while injecting new vitality into the digital transformation of various industries. For more information, visit the platform’s official website: [https://platform.lingyiwanwu.com/](https://platform.lingyiwanwu.com/).
【来源】https://mp.weixin.qq.com/s/qmjEe56WtQeUgV5iyTx4pg
Views: 2