Zhipu AI: Pioneering Multilingual AI from Tsinghua’s Halls
Born from the academic rigor of Tsinghua University in 2019, Zhipu AI stands out as a pioneer in the realm of bilingual artificial intelligence within China. Founded by two distinguished professors, the company has quickly established itself as a key player, boasting an impressive portfolio of AI solutions. Zhipu AI’s journey began with a vision to create AI models that could understand and generate content in both Chinese and English, catering to the growing need for cross-lingual communication and information access. This focus on bilingual capabilities has set Zhipu AI apart from many other AI startups in China, which often prioritize the domestic market.
At the heart of Zhipu AI lies ChatGLM, a sophisticated chatbot designed to engage users in natural and intuitive conversations. ChatGLM is more than just a chatbot; it’s a versatile AI assistant capable of handling a wide range of tasks, from answering general knowledge questions to providing personalized recommendations. Complementing this conversational AI is Ying, an innovative AI-powered video generation tool that empowers users to create compelling visual content with ease. Ying democratizes video creation, allowing users with little to no video editing experience to produce professional-looking videos with just a few clicks. This combination of conversational and visual AI tools positions Zhipu AI as a comprehensive AI solution provider.
Last August marked a significant milestone for Zhipu AI with the unveiling of its GLM-4-Plus model. This advanced AI model has garnered widespread acclaim for its performance, drawing comparisons to OpenAI’s GPT-4o. The GLM-4-Plus model is a testament to Zhipu AI’s commitment to pushing the boundaries of AI technology. Zhipu AI further expanded its capabilities with the introduction of GLM-4-Voice, a conversational AI model capable of speaking both Mandarin Chinese and English with remarkable nuance. GLM-4-Voice opens up new possibilities for voice-based AI applications, such as virtual assistants, language learning tools, and accessibility solutions.
Despite these achievements, Zhipu AI recently faced a setback when the U.S. government added the company to its trade restriction list. This highlights the geopolitical complexities surrounding AI development and the potential for government intervention. However, this obstacle did not deter investors. Earlier this month, Zhipu AI successfully raised over $140 million, with participation from Alibaba, Tencent, and several state-backed funds. This funding round demonstrates the continued confidence in Zhipu AI’s potential and its importance to China’s AI strategy.
Moonshot AI: Revolutionizing Conversational AI with Kimi
Emerging from the same prestigious halls of Tsinghua University in 2023, Moonshot AI is the brainchild of Yang Zhilin, a distinguished researcher with a background from Carnegie Mellon. Yang Zhilin’s experience at Carnegie Mellon, a leading AI research institution, has undoubtedly influenced Moonshot AI’s approach to developing cutting-edge AI technology. The company’s flagship product, the Kimi AI chatbot, has rapidly gained popularity, securing its place among the top 5 most widely used chatbots in China. Kimi AI’s success is attributed to its user-friendly interface, its ability to handle complex queries, and its continuous improvement based on user feedback.
As of November 2023, Kimi AI boasted nearly 13 million monthly active users, according to Counterpoint Research. This impressive user base underscores Kimi AI’s appeal and its relevance to the Chinese market. The chatbot’s standout feature is its ability to process queries of up to 2 million Chinese characters, enabling users to engage in in-depth conversations and complex information retrieval. This long-context processing capability sets Kimi AI apart from many other chatbots, which often struggle to handle lengthy conversations or complex documents. Kimi AI’s ability to understand and respond to long-form text makes it a valuable tool for research, writing, and information analysis.
Valued at $3.3 billion, Moonshot AI benefits from the backing of industry giants Alibaba and Tencent. These strategic partnerships provide Moonshot AI with access to resources, expertise, and distribution channels that can accelerate its growth. The company’s rapid growth and innovative chatbot position it as a rising star in China’s AI landscape. Moonshot AI’s future plans include expanding Kimi AI’s capabilities, developing new AI-powered applications, and entering new markets.
MiniMax: Crafting Virtual Worlds with AI-Powered Characters
Founded in 2021 by AI researcher Yan Junjie, MiniMax distinguishes itself by focusing on the creation of AI-powered virtual characters. Yan Junjie’s background in AI research has enabled MiniMax to develop sophisticated algorithms for creating realistic and engaging virtual characters. The company’s flagship product, Talkie, allows users to engage in conversations with a diverse cast of virtual personalities, ranging from celebrities to fictional characters. Talkie offers a unique and immersive experience, allowing users to interact with AI-powered characters in a way that feels natural and authentic.
Originally launched in 2022 as Glow, the app was later rebranded as Xingye in China and Talkie in international markets. This rebranding reflects MiniMax’s ambition to expand its reach beyond China and to appeal to a global audience. However, Talkie was removed from the U.S. App Store in December due to ‘technical reasons,’ according to the South China Morning Post. This setback highlights the challenges of navigating the complex regulatory landscape and technical requirements of different app stores.
In addition to Talkie, MiniMax has developed Hailuo AI, a text-to-video tool that allows users to create videos from written content. Hailuo AI complements Talkie by providing users with another creative outlet for expressing themselves and sharing their stories. The company is valued at $2.5 billion following a $600 million funding round led by Alibaba in March last year. This funding round demonstrates the strong investor interest in MiniMax’s vision and its potential to disrupt the entertainment industry.
Baichuan Intelligence: Building Open-Source AI for All
Established in March 2023, Baichuan Intelligence brings together talent from Microsoft, Huawei, Baidu, and Tencent. This diverse team of experts brings a wealth of experience and expertise to Baichuan Intelligence, enabling the company to develop innovative AI solutions. The company has made significant contributions to the open-source AI community by releasing two open-source language models: Baichuan-7B and Baichuan-13B. Baichuan Intelligence’s commitment to open-source AI reflects its belief that AI should be accessible to everyone and that collaboration is essential for driving innovation.
These models are trained on multilingual data and support a wide range of applications, including general knowledge, mathematics, programming, translation, law, and medicine. Baichuan-7B and Baichuan-13B are designed to be versatile and adaptable, making them suitable for a wide range of tasks and industries. In July, Baichuan raised $687.6 million, valuing the company at nearly $2.8 billion. This substantial funding round underscores the importance of open-source AI and the potential for Baichuan Intelligence to become a major player in the global AI ecosystem.
StepFun: Pushing the Boundaries of Foundation Models
Founded in 2023 by Jiang Daxin, former Senior Vice President of Microsoft, StepFun has quickly made a name for itself in the AI industry. Jiang Daxin’s experience at Microsoft, a leading technology company, has given him a deep understanding of the challenges and opportunities in the AI field. Despite being a relatively young company, StepFun has launched 11 foundation models, including AI models for image processing, audio processing, and multimodal applications. StepFun’s rapid development and deployment of foundation models demonstrate its agility and its commitment to innovation.
StepFun’s most notable creation is Step-2, a language model with one trillion parameters. Step-2 is one of the largest and most powerful language models in the world, capable of generating highly realistic and coherent text. This model is ranked among the top performers on the LiveBench leaderboard, competing with models from DeepSeek, Alibaba, and OpenAI. StepFun’s success in developing Step-2 demonstrates its technical prowess and its ability to compete with the world’s leading AI companies. Last December, StepFun raised hundreds of millions of dollars in a Series B funding round. This funding will enable StepFun to continue developing its foundation models and to expand its team and infrastructure.
01.AI: Championing Open-Source AI with a Human Touch
Founded in 2023 by Kai-Fu Lee, 01.AI has emerged as a prominent player in China’s open-source AI movement. Kai-Fu Lee’s extensive experience in the AI field, including his leadership roles at Google and Microsoft, has given him a unique perspective on the potential of AI and the importance of open-source collaboration. The company’s two main models, Yi-Lightning and Yi-Large, have quickly gained recognition for their performance in language understanding, reasoning, and contextual awareness. Yi-Lightning and Yi-Large are designed to be accessible and easy to use, making them suitable for a wide range of applications and users.
Yi-Lightning stands out for its cost-effective training. According to Kai-Fu Lee, the model was trained using only 2,000 Nvidia H100 GPUs for one month, significantly less than xAI’s Grok 2. This demonstrates 01.AI’s ability to develop high-performance AI models with limited resources. Yi-Large is designed for natural human-like conversations, supporting both Chinese and English. Yi-Large’s focus on natural language understanding and generation makes it a valuable tool for developing chatbots, virtual assistants, and other conversational AI applications.
In summary, while DeepSeek captures headlines, China’s AI advancement is propelled by the collective innovation and strategic prowess of the ‘Six Tigers.’ Their diverse strengths, from multilingual AI to virtual characters and open-source models, underscore the depth and breadth of China’s AI capabilities. These companies represent the cutting edge of AI development in China, and their contributions are shaping the future of the technology. As these companies continue to evolve and compete, they will undoubtedly shape the future of AI, not just in China, but on a global scale. The ‘Six Tigers’ are driving innovation, fostering collaboration, and pushing the boundaries of what’s possible with AI, and their impact will be felt for years to come. Their success is a testament to the talent, ambition, and strategic vision of China’s AI community.