Alibaba upgrades AI model Tongyi Qianwen, releases industry-specific models

HONG KONG, Oct 31 (Reuters) – Chinese technology giant Alibaba (9988.HK) on Tuesday updated its artificial intelligence (AI) model Tongyi Qianwen with industry-specific announced the release of an AI model suite. companies.

Alibaba’s cloud computing arm said at its annual conference in Hangzhou that Tongyi Qianwen 2.0 has “hundreds of billions” of parameters (a benchmark used to measure the ability of AI models) and that it ranks among the world’s best in its metrics. He said it has become one of the most powerful AI models.

The company also announced eight AI models for the entertainment, finance, healthcare, and legal industries.

The upgrade comes just six months after the model’s initial release and reflects the pace at which tech companies are racing to control China’s nascent and rapidly expanding AI market.

Rival Tencent (0700.HK) said last month that China was in the midst of a “100 model war”, with more than 130 models flooding the market. Tencent said the company’s Hunyuan AI has over 100 billion parameters and outperforms OpenAI’s GPT-4 model in processing Chinese.

OpenAI’s GPT-3 AI model, backed by Microsoft (MSFT.O), will contain 175 billion parameters in 2020, and Meta Platform’s (META.O) Llama 2 model will contain 70 billion parameters in 2023. It contained parameters.

Alibaba’s industry-specific model went online on Tuesday, offering specialized tools for creating images, writing computer code, analyzing financial data and searching legal documents.

Chairman Joe Tsai also said at the conference that about half of China’s large-scale language AI models are currently running on Alibaba Cloud.

He said Alibaba’s AI model sharing platform ModelScope currently has 2,300 models and 2.7 million developers contributing.

Our standards: Thomson Reuters Trust Principles.