Beijing, China – Wednesday, 29 January 2025 – While millions in China enjoyed Lunar New Year with their families, tech giant Alibaba made a bold move to stay ahead in the country’s fiercely competitive AI race. The company released an upgraded version of its Qwen 2.5 AI model that it claimed now surpassed the highly-acclaimed DeepSeek-V3.
The DeepSeek-V3, which is DeepSeek's latest AI model, is a fast-growing player in China’s AI industry, shaking up both local competitors and Silicon Valley. Recently, DeepSeek has just claimed the most download app on the United States's Apple App Store and Google Play Store, passing 5M download globally on Google Play Store.
Why the rush?
Alibaba’s timing raised eyebrows. Releasing Qwen 2.5-Max on the first day of Lunar New Year – a time when most workers are on holiday – hints at the pressure Chinese tech firms face after DeepSeek’s explosive rise.
Since mid-January, the startup has released two powerful AI models, DeepSeek-V3 and DeepSeek R1. These models are praised for being low-cost while competing with OpenAI’s GPT-4. Their success even affected US markets, causing tech stocks to fall as investors worried about the high costs of American AI projects.
However, Alibaba’s cloud division boldly stated on WeChat that its new model “outperforms almost across the board” top rivals like GPT-4o and Meta’s Llama-3.1-405B. But critics note Qwen 2.5 still lacks features like ChatGPT’s voice interaction tools.
Well, DeepSeek's impact is undeniable. In May 2023, DeepSeek’s earlier model, V2, started a price war by offering AI services for just 1 yuan (£0.11) per million tokens—much cheaper than other companies. In response, Alibaba cut prices by 97%, and Baidu and Tencent did the same.
Yet DeepSeek’s founder, Liang Wenfeng, shrugged off the competition, telling Chinese media his team is laser-focused on developing artificial general intelligence (AGI) – AI that surpasses humans in key tasks.
Unlike Alibaba’s vast workforce, DeepSeek runs like a nimble research lab, staffed by top university graduates. Liang argues tech giants’ “high costs and rigid structures” could hold back AI innovation – a dig at firms like Alibaba.
US vs China: The Global AI Battle
The rivalry isn’t just domestic. As DeepSeek gained traction, OpenAI’s Sam Altman teased “exciting” ChatGPT updates, while Microsoft’s Satya Nadella hinted at new AI projects.
Things got tense when OpenAI accused DeepSeek of secretly using ChatGPT’s data to train its AI. If true, this would break OpenAI’s rules. But ironically, OpenAI has also been accused of using copyrighted content without permission.
What’s Next?
With ByteDance (TikTok’s owner) also boosting its AI model days after DeepSeek’s R1 launch, China’s tech titans are in a sprint. Alibaba’s holiday release shows even giants can’t afford to slow down. As costs plunge and capabilities grow, one thing’s clear: the AI race is heating up on all fronts – and everyone’s scrambling to keep up.
