Alibaba claims its AI model is better than DeepSeek

Alibaba.png

Since the launch of the DeepSeek-V3 model on January 10 and the R1 model on January 20, the tech world, including Silicon Valley, has been stunned by DeepSeek's supposedly cheaper and more effective capabilities.  In response, local competitors in China raced to upgrade their AI models. Two days after DeepSeek-R1 was introduced, ByteDance, the owner of TikTok, updated its main AI model and claimed it outperformed OpenAI's o1 model in AIME tests.

Previously, the launch of the DeepSeek-V2 model sparked a price war in the AI industry in China. The model is open-source and cheap—only 1 yuan (~RM0.61) for every 1 million tokens. This forced Alibaba to lower prices by up to 97% for some of their models, followed by other giants such as Baidu and Tencent. DeepSeek operates differently from the Chinese tech giants. It operates like a research lab, with a workforce made up of young graduates and PhD holders from top universities in China. Liang Wenfeng, the founder of DeepSeek, once stated that his company doesn’t care about price wars because its ultimate goal is to develop Artificial General Intelligence (AGI)—an AI that can outperform humans at economically valuable tasks.

He also believes that large tech companies like Alibaba may not be suitable for the future of the AI industry because of their overly complex structure and high operating costs. "Large AI models require continuous innovation, and the capabilities of giant tech companies are limited," Liang said in an interview. With AI competition intensifying in China, Alibaba, DeepSeek, ByteDance, and others continue to race to introduce more sophisticated and cheaper technology. DeepSeek's success shows that lighter, cheaper, and more innovative AI models may be the key to the future of the global AI industry.

Furthermore, this puts pressure on the United States, which has previously imposed new restrictions on the export of artificial intelligence chips aimed at maintaining the US's dominant position in AI development. We have yet to confirm other details on the news. Hopefully, everything will go smoothly.

 


What are your thoughts on this news? Comment down below and stay tuned for more news like this at TechNave!