MEITUAN-WR Steps Up AI Efforts: Invests in Prominent AI Modeling Company Zhipu AI, Holding Over 10% Stake

After 360, MEITUAN-WR also sees the potential of Zhifu AI.

According to Tianyancha App, Beijing Zhifu Huazhang Technology Co., Ltd., an affiliated company of Zhifu AI, recently underwent a business change. Its registered capital increased from approximately 14.8069 million RMB to approximately 16.5286 million RMB. MEITUAN-WR's subsidiary, Tianjin Sankuai Technology Co., Ltd., became a new shareholder with a 10.42% stake.

Beijing Zhifu Huazhang Technology Co., Ltd. was established in June 2019, with Liu Debing as the legal representative. Its business scope includes technology services, development of artificial intelligence software, data processing and storage support services, and technology intermediary services.

As previously reported by Wall Street News, Zhifu AI was founded by Professor Tang Jie from the Department of Computer Science at Tsinghua University. The core members of the team have participated in the development of the "Wudao" project, a collaboration between Tsinghua University and the Institute of Automation, Chinese Academy of Sciences. In August 2022, the Knowledge Engineering Laboratory of Tsinghua University and Zhifu AI jointly developed the large-scale pre-training language model GLM-130B, which was officially launched. It is the only Asian large-scale model selected for Stanford evaluation in the same year.

GLM has performed exceptionally well in key indicators such as accuracy compared to large models from companies such as OpenAI, Google Brain, and Meta. It has surpassed GPT-3, Google's PaLM, and Meta's OPT large models in MMLU, LAMBADA, and BIG-bench-lite benchmark tests.

On May 16th of this year, 360 announced a strategic partnership with Zhifu AI to jointly develop a trillion-level large model called "360GLM". The two parties will refer to the cooperation model of "Microsoft + OpenAI" and combine large models with application scenarios.

Zhou Hongyi, CEO of 360, believes that China should establish a collaborative innovation model between large technology companies and key research institutions to lead the technological breakthroughs in large models. He stated that this cooperation with Zhifu AI is precisely based on this collaborative relationship between industry and research. Regarding this collaboration, Zhang Peng, the CEO of Zhifu AI, expressed that Zhifu AI has always adhered to its own vision of making machines think like humans and realizing the concept of Model as a Service (MaaS).

Currently, the training data volume of this model is 400 billion, with Chinese and English each accounting for half. It has 130 billion parameters and the training cost ranges from millions to tens of millions of RMB. As of May 1st this year, the model has received more than 1,000 download and usage requests from 69 countries and over 1,000 research institutions.

Based on GLM-130B, Zhifu AI has further fine-tuned and developed the ChatGLM model with supervised learning. ChatGLM is currently the most advanced open-source large-scale model in China and has been made available for internal testing.

Can Zhifu AI become China's next OpenAI?