Alibaba’s Qwen 2.5 Surpasses DeepSeek-V3, Escalating China’s AI Rivalry
Chinese tech giant Alibaba Group Holding Ltd. has unveiled its latest AI model, Qwen 2.5-Max, asserting it outperforms DeepSeek-V3, which has gained significant attention recently.
The Qwen 2.5 announcement, via the Taipei Times, made on the first day of the Lunar New Year, underscores the intensifying competition in China’s AI sector. DeepSeek, a rising AI start-up, has disrupted the landscape with its rapid advancements, pressuring established players like Alibaba, ByteDance, and Tencent to accelerate their own AI developments.
Qwen 2.5 vs. DeepSeek-V3: A New Benchmark
Alibaba’s cloud division stated in an official WeChat post that Qwen 2.5-Max “outperforms … almost across the board GPT-4o, DeepSeek-V3, and Llama-3.1–405B.” This claim positions the model as a formidable competitor not only within China but against leading AI models from OpenAI and Meta.
DeepSeek’s AI assistant, powered by the DeepSeek-V3 model, launched on Jan. 10, followed by the R1 model on Jan. 20. These releases reportedly shocked Silicon Valley, causing turbulence in tech stocks.
DeepSeek’s ability to deliver high performance at a fraction of traditional AI costs has raised concerns among investors about the sustainability of AI development expenditures in the U.S.
China’s AI Battle Heats Up
DeepSeek’s advancements have sparked a flurry of activity among domestic competitors. Two days after DeepSeek-R1’s release, ByteDance Ltd. introduced an upgrade to its flagship AI model, claiming it outperformed OpenAI’s o1 in AIME, a benchmark test for complex task comprehension.
This echoes DeepSeek’s own assertion that R1 rivaled OpenAI’s o1 across multiple performance metrics. The trend highlights a broader push among Chinese firms to challenge global AI leaders while maintaining cost efficiency.
A History of Disruption
DeepSeek’s influence on the AI sector dates back to its May 2023 launch of DeepSeek-V2. Its open-source model, priced at just 1 yuan (US$0.14) per 1 million tokens, triggered a pricing war among China’s tech giants. In response, Alibaba slashed costs on its own AI offerings by up to 97%, prompting similar moves from Baidu and Tencent Holdings Ltd.
Despite the aggressive competition, DeepSeek’s enigmatic founder, Liang Wenfeng, remains unfazed. In a rare July 2023 interview, he stated that DeepSeek’s primary objective is achieving artificial general intelligence, not engaging in price battles.
“Large foundational models require continued innovation; tech giants’ capabilities have their limits,” Liang remarked, contrasting DeepSeek’s lean approach with the high-cost structures of larger corporations.
The Future of AI in China
With Alibaba’s latest move, China’s AI race is expected to intensify further. As tech giants and emerging players push innovation boundaries, the country’s AI sector is rapidly evolving into a globally competitive force.
