(Yicai) Aug. 31 -- China's first eight large language models, including those of Baidu, SenseTime, and ByteDance, have been approved to open to the public today.

The eight LLMs include Baidu's Ernie Bot, SenseTime's SenseChat, Zhipu AI's GLM, ByteDance's Yunque, Zidong Taichu of the Chinese Academy of Sciences, MiniMax’s ABAB, and Shanghai Artificial Intelligence Laboratory's Intern.

Users can already register for free to use such ChatGPT-like programs developed by Baidu, SenseTime, Tsinghua University-backed Zhipu AI, and Baichuan Intelligent Technology.

The group might get bigger soon as Alibaba has applied for a similar permit for its Tongyi Qianwen, as well as iFlytek for its Spark Model. So far almost 80 LLMs with more than one billion parameters have been launched in China but before obtaining regulatory approvals, the models can only be tested by a small group of users.

The approvals should increase user access and speed up model iteration, making the LLMs smarter, according to an insider. Next, market players are expected to intensify rivalry in the application field.

As new opportunities emerge, new regulations follow. On Aug. 15, seven bodies, including the internet watchdog, issued new rules to require generative AI providers to assess security, report algorithm changes, and enroll in those models.

The McKinsey Global Institute estimated in a report in June that generative AI could add from USD2.6 trillion to USD4.4 trillion in value per year across the 63 use cases they analyzed.

Editor: Emmi Laine