Chinese tech giant Alibaba (9988.HK) unveiled a new version of its Qwen 2.5 artificial intelligence model on Wednesday, claiming it surpasses the highly acclaimed DeepSeek-V3. The release of Qwen 2.5-Max on the first day of the Lunar New Year, a time when most Chinese citizens are on holiday with their families, underscores the intense pressure Chinese AI startup DeepSeek has exerted on both domestic and international competitors following its rapid rise over the past three weeks.
In a statement posted on its official WeChat account, Alibaba’s cloud unit asserted that “Qwen 2.5-Max outperforms… almost across the board GPT-4o, DeepSeek-V3, and Llama-3.1-405B,” referencing the most advanced open-source AI models from OpenAI and Meta.
DeepSeek’s recent releases—the DeepSeek-V3-powered AI assistant on January 10 and the R1 model on January 20—have sent shockwaves through Silicon Valley, causing tech stocks to plummet. The startup’s reportedly low development and operational costs have raised questions among investors about the massive spending plans of leading U.S. AI firms. Meanwhile, DeepSeek’s success has spurred its domestic rivals to accelerate upgrades to their own AI models.
Just two days after DeepSeek-R1’s launch, ByteDance, the owner of TikTok, rolled out an update to its flagship AI model, claiming it outperformed Microsoft-backed OpenAI’s o1 in the AIME benchmark test, which evaluates AI models’ ability to understand and respond to complex instructions. This mirrored DeepSeek’s assertion that its R1 model rivals OpenAI’s o1 on multiple performance benchmarks.
DeepSeek’s earlier model, DeepSeek-V2, sparked an AI price war in China after its release in May last year. Its open-source nature and unprecedented affordability—charging just 1 yuan ($0.14) per 1 million tokens (units of data processed by the AI model)—prompted Alibaba’s cloud unit to slash prices by up to 97% on a range of its models. Other Chinese tech giants, including Baidu (9888.HK) and Tencent (0700.HK), quickly followed suit.
In a rare interview with Chinese media outlet Waves in July, DeepSeek’s enigmatic founder, Liang Wenfeng, stated that the startup was unfazed by price wars, emphasizing that its primary goal was achieving AGI (artificial general intelligence). OpenAI defines AGI as autonomous systems that outperform humans in most economically valuable tasks.
Unlike large Chinese tech companies such as Alibaba, which employ hundreds of thousands of workers, DeepSeek operates more like a research lab, primarily staffed by young graduates and doctoral students from China’s top universities. Liang highlighted the contrast between the high costs and rigid structures of tech giants and DeepSeek’s lean operations and flexible management style, suggesting that the former might struggle to adapt to the future of the AI industry.
“Large foundational models require continued innovation; tech giants’ capabilities have their limits,” Liang remarked.
Reporting by Eduardo Baptista; Editing by Christian Schmollinger.
