Capable of conversing in both Chinese and English, Tencent’s large language model ‘Hunyuan’ is claimed to be better than OpenAI’s ChatGPT in several areas
After getting AI public release approvals from Chinese Authorities last month, Tencent has finally released its large language artificial intelligence (AI) model ‘Hunyuan’, making its entry into the competitive global AI race.
Aiming to become China’s AI leader, the WeChat parent firm launched the ‘Hunyuan’ AI model in front of a large audience at a conference in Shenzhen. Tencent officials, during the launch mentioned how its AI model is becoming the foundation for more than 50 of its products and services.
“By July, there are more than 130 large language models in China,” said Tencent’s vice president, Jiang Jie. “A war of a hundred models has begun,” he added.
Now available to use for enterprises, Tencent’s Hunyuan comes at a time where several other Chinese tech firms such as Baidu Inc and SenseTime Group have also released their own large language models and AI products, thus creating a wave of competition.
According to details, Tencent has trained Hunyuan with more than 2 trillion tokens, giving the large language model over 100 billion parameters.
For comparison, OpenAI’s former AI model GPT-3 had nearly 175 billion parameters whereas Meta’s latest Llama 2 model has over 70 billion parameters.
Capable of conversing in both Chinese and English, Tencent’s large language model ‘Hunyuan’ is claimed by Tencent to be better than OpenAI’s ChatGPT in several areas such as writing long text or solving certain math problems.
Tencent has also made several other claims related to hallucinations, but none of them including the ones that include its comparison with ChatGPT are independently verified.