Capable of conversing in both Chinese and English, Tencent’s large language model ‘Hunyuan’ is said to outperform OpenAI’s ChatGPT in various aspects
After gaining approval for public release from Chinese authorities last month, Tencent has officially introduced its large language artificial intelligence (AI) model ‘Hunyuan’, making its mark in the globally competitive AI landscape.
Striving to lead in China’s AI sector, the parent company of WeChat showcased the ‘Hunyuan’ AI model to a significant audience at a conference in Shenzhen. During the launch, Tencent representatives highlighted how the AI model serves as the foundation for more than 50 of its products and services.
Tencent’s vice president, Jiang Jie, stated, “As of July, there are over 130 large language models in China,” adding, “A competition of a hundred models has commenced.”
Now available for enterprise use, Tencent’s Hunyuan enters the market at a time when several other Chinese tech firms like Baidu Inc and SenseTime Group have also unveiled their own large language models and AI products, intensifying the competition.
Tencent reportedly trained Hunyuan using more than 2 trillion tokens, equipping the large language model with over 100 billion parameters.
For comparison, OpenAI’s previous AI model GPT-3 had nearly 175 billion parameters, while Meta’s latest Llama 2 model has over 70 billion parameters.
Capable of conversing in both Chinese and English, Tencent asserts that its large language model ‘Hunyuan’ exceeds OpenAI’s ChatGPT in various aspects such as generating long text or solving specific mathematical problems.
Tencent has also made several other claims related to hallucinations, but none of them, including the comparisons with ChatGPT, have been independently verified.
Image Source:Tada Images @ShutterStock
