In an unexpected move on the first day of Lunar New Year, Chinese tech giant Alibaba announced its latest AI model, Qwen 2.5-Max, claiming it outperforms several leading models including the recently released DeepSeek-V3.
The unusual timing of the announcement, when most Chinese people are celebrating with their families, highlights the growing pressure from DeepSeek’s rapid rise in the AI landscape.
Alibaba Fires Back in China’s AI Battle After DeepSeek’s Surprise Success
Alibaba’s cloud division stated that their new model “Qwen 2.5-Max outperforms … almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B.”
This announcement comes in response to DeepSeek’s recent achievements that have caught the attention of both Chinese and international tech communities.
DeepSeek, a relatively small Chinese AI startup, has been making waves since January. The company’s AI assistant powered by DeepSeek-V3, launched on January 10, followed by their R1 model on January 20, has created significant ripples in Silicon Valley.
The startup’s ability to develop high-performing AI models at surprisingly low costs has raised questions about the massive spending by leading U.S. AI companies.
The impact of DeepSeek’s success is evident in the Chinese tech sector. ByteDance, the company behind TikTok, quickly responded with an update to its own AI model just two days after DeepSeek-R1’s release. ByteDance claimed its model performed better than Microsoft-backed OpenAI’s o1 in the AIME benchmark test, which measures AI models’ ability to understand and respond to complex instructions.
DeepSeek’s Disruptive Entry
DeepSeek’s influence on the Chinese AI market began even earlier. Last May, their V2 model triggered a price war in China by offering unprecedented low rates – just “1 yuan ($0.14) per 1 million tokens.” This aggressive pricing strategy forced Alibaba to slash its AI model prices by up to 97%, with other major players like Baidu and Tencent following suit.
What makes DeepSeek particularly interesting is its unique approach to AI development. Unlike traditional Chinese tech giants with their massive workforces, DeepSeek operates more like a research lab. The company primarily employs young graduates and doctorate students from top Chinese universities, maintaining a lean operation with minimal management hierarchy.
Can Agile Startups Outpace Tech Giants in the AI Race?
DeepSeek’s founder, Liang Wenfeng, has shown little interest in price competition. In a rare interview with Chinese media outlet Waves, Liang emphasized that his company’s main focus is achieving artificial general intelligence (AGI) – defined by OpenAI as autonomous systems that can outperform humans in most economically valuable tasks.
Liang’s perspective on the future of AI development challenges traditional assumptions about big tech companies’ advantages. He suggested that major tech companies might actually be at a disadvantage in the AI race, citing their high operational costs and rigid organizational structures as potential limitations.
“Large foundational models require continued innovation, tech giants’ capabilities have their limits,” Liang stated, highlighting his belief in the advantages of smaller, more agile organizations in pushing AI technology forward.
This ongoing competition in China’s AI sector shows how quickly the landscape is evolving, with smaller, focused companies like DeepSeek challenging established tech giants through innovation and efficient operations. The race to develop more advanced AI models continues to intensify, potentially reshaping the future of both Chinese and global AI development.
