Chinese IT giant Tencent has la its proprietary large foundation model, Hunyuan, making it available for enterprise use.
The Hunyuan large language model is designed for enterprises in China to test and build applications via APIs on Tencent Cloud. The model offers strong Chinese language processing abilities, advanced logical reasoning, and reliable task execution capabilities. Tencent's foundation model can be utilized in various functions, including image creation, copywriting, text recognition, and customer service. These features are expected to be pivotal in industries such as finance, public services, e-commerce, transportation, and gaming.
Comparative Analysis with Other Models
According to a report by Computerworld, Hunyuan is positioned as a powerful and intelligent model, surpassing US-developed models like OpenAI ChatGPT and Meta's Llama-2. The model boasts over 100 billion parameters and was trained with more than 2 trillion tokens. In comparison, OpenAI's GPT-3 model contained 175 billion parameters when launched in 2020, and Meta's Llama-2 model had 70 billion parameters upon its release.
Integration with Tencent's Offerings
Tencent has integrated Hunyuan with more than 50 of its products and services, including Tencent Cloud, Tencent Marketing Solutions, Tencent Games, and the WeChat social media app. For instance, Tencent Meeting now features a Hunyuan-powered AI assistant that can generate relevant meeting minutes by processing natural language instructions from users. Additionally, Tencent Docs, integrated with Hunyuan, supports various text creation scenarios and can generate standard format text with a single click.
Dowson Tong, Senior Executive Vice President of Tencent and CEO of Tencent Cloud and Smart Industries Group (CSIG), stated at the Global Digital Ecosystem Summit in Shenzhen that “…in launching Hunyuan and making it available to domestic enterprises, Tencent has opted for an approach that balances the exciting performance of consumer-facing, large-model AI powered chatbots, with the pragmatic need for the business community to increase operational efficiencies, reduce costs, protect privacy as well as proprietary data.”
Jie Jiang, Vice President of Tencent, emphasized the company's evolution over the years, noting that Tencent has transitioned from model algorithms to machine learning frameworks and now, to AI infrastructure. He highlighted that “…the Hunyuan model has over 100 billion parameters, with more than two trillion tokens in pre-training data.”
Regulatory Landscape in China
China has been proactive in regulating AI. The Cyberspace Administration of China introduced measures for managing generative AI services, making providers accountable for the validity of data used to train generative AI tools. Under these regulations, AI firms are required to submit security assessments to the government before launching AI tools to the public.
The first batch of AI chatbots has recently been approved by China for public use, which is a major milestone for the country's growing AI industry. Some of the big names that are offering their generative AI services to the public are Baidu Inc. and ByteDance Ltd. This comes after Beijing gave the green light to some companies to launch their AI products, an area where China has poured billions since OpenAI ChatGPT came out.
Baidu's Ernie Bot has become a hit, making it the top free app in China on the Apple App Store. In just one day after it was released, Ernie Bot got over a million users, according to the company. There are also other chatbot apps that have been made by startups backed by major players like Tencent, Meituan, and famous investors.
