Starting from August 31, 11 domestic AI large model products registered through the "Interim Measures for the Management of Generative Artificial Intelligence Services" will be launched one after another and open to the whole society. The five large-scale model products in Beijing are Baidu’s “Wenxinyiyan”, Douyin’s “Skylark”, Baichuan Intelligent’s “Baichuan Large Model”, Tsinghua-based AI company Zhipu Huazhang’s “Zhipu Qingyan” and the Chinese Academy of Sciences "Zidong Taichu". The three large-scale model products in Shanghai include SenseTime’s “SenseChat”, MiniMax’s “ABAB Large Model”, and Shanghai Artificial Intelligence Laboratory’s “Scholar’s Universal Large Model”. Let us take a look at the profiles of the first eight large model companies approved to go online.
1. Baidu’s “Wen Xin Yi Yan” will be fully open to the whole society on August 31. Users can download the "Wenxin Yiyan APP" from the app store or log in to the "Wenxin Yiyan Official Website" (https://yiyan.baidu.com) to experience it. Baidu will also open a batch of newly reconstructed AI native applications, allowing users to experience the four core capabilities of generative AI: understanding, generation, logic, and memory. 
Wen Xinyiyan opened the invitation test on March 16th. Robin Li, chairman and CEO of Baidu, said that when Wen Xinyiyan opens its services to hundreds of millions of Internet users on a large scale, it will be able to obtain a large amount of real-world manual feedback, which will further improve the basic model and speed up the development of the service. Iterate Wenxinyiyan to create a better user experience.
2. Baichuan Intelligence On August 31, Baichuan Intelligence, founded by former Sogou CEO Wang Xiaochuan, announced that it was the first to pass the "Interim Measures for the Management of Generative Artificial Intelligence Services" and will open its services to the whole society from today. Users can log in to www.baichuan-ai.com to experience the new experience brought by Baichuan Big Model in areas such as knowledge Q&A and text creation. Only four months after its establishment, Baichuan Intelligence has released three general-purpose large language models. Baichuan Intelligent said that as of today, the downloads of the two open source models Baichuan-7B/13B on Huggingface have exceeded three million.
3. SenseTime’s large model application “SenseChat” SenseTime’s large model application “SenseChat” was also announced on August 31 and will be fully open to users from now on. Users can experience it through the URL https://chat.sensetime.com. After the user's mobile phone is registered, he or she can have a conversation with "SenseChat". On April 10, SenseTime launched the large-scale model "RiRixin", including the natural language processing model "Consultation", the Vincent graph model "Miahua" and the digital human video generation platform "Ruying". Among them, SenseChat 1.0 is also one of the earliest chat robot products launched in China based on a large language model with hundreds of billions of parameters.
4. "Zhipu Qingyan" (based on ChatGLM2) by Tsinghua-based AI company Zhipu Huazhang. On August 31, Zhipu Huazhang officially launched the first generative AI assistant - "Zhipu Qingyan", which is now available on All major app stores are online. Users can download it through the Apple App Store, mainstream Android stores (including Huawei, OPPO, vivo and Xiaomi, etc.), or search for "Zhipu Qingyan" in the WeChat mini program to experience its functions. 
The assistant is based on the Chinese-English bilingual dialogue model ChatGLM2 independently developed by Zhipu AI. It has been pre-trained with trillions of characters of text and code, and uses supervised fine-tuning technology to provide users with intelligent services in the form of universal dialogue. As an intelligent assistant for users, "Zhipu Qingyan" can answer various questions and complete various tasks for users in work, study and daily life. At present, "Zhipu Qingyan" already has rich capabilities such as "general question and answer, multi-round dialogue, creative writing, code generation and virtual dialogue", and will also open up multi-modal and other generation capabilities in the future.
5. Douyin’s Skylark model In mid-August, Doubao, an AI robot developed based on the Skylark model of Douyin, a subsidiary of ByteDance, began a small-scale invitation test. Users can log in through their mobile phone number, TikTok or Apple ID. "Doubao" provides functions such as chatbots, writing assistants, and English learning assistants. It can answer various questions and conduct conversations to help people obtain information. 
6. “Zidong Taichu” under the Chinese Academy of Sciences “Zidong Taichu” will be open to the public on September 1. "Zidong Taichu" is positioned as a cross-modal general artificial intelligence platform and will be officially released in 2021. It has multiple modes of vision, text, and voice, and is built on a full-stack localized basic software and hardware platform. In June, "Zidong Taichu" released version 2.0, adding new modal data such as video, sensor signals, and 3D point clouds. At present, "Zidong Taichu" has been applied in many industries such as text creation, medical care, logistics, and government affairs.
7. The scholar model of the Shanghai Artificial Intelligence Laboratory The Shanghai Artificial Intelligence Laboratory was established in 2020 and was initiated and established by the Shanghai Artificial Intelligence Innovation Center. In 2021, the Shanghai Artificial Intelligence Laboratory, together with SenseTime, the Chinese University of Hong Kong, and Shanghai Jiao Tong University, released the general vision technology system "Scholar" (INTERN). Today, the parameters of the Shusheng large model system have reached 100 billion levels, including three basic models: "Shusheng·Multimodal", "Shusheng·Puyu" and "Shusheng·Tianji", as well as for the development and application of large models. Open source system.
8. MiniMax’s ABAB large model MiniMax was founded in December 2021 by Yan Junjie, the former vice president of SenseTime and head of general intelligent technology, and is known as All in AGI. ABAB is an advanced general-purpose large language model with powerful language processing capabilities that can understand and generate natural language text.