Chinese tech giant Alibaba has launched an upgraded version of its AI model, Qwen 2.5-Max, claiming it surpasses leading models such as DeepSeek-V3, GPT-4o, and Llama-3.1-405B. The timing of the release—on the first day of the Lunar New Year—reflects the urgency Chinese companies feel in keeping pace with the rapid rise of AI startup DeepSeek.
DeepSeek's AI models have sent shockwaves through the tech industry, particularly in Silicon Valley, where its low-cost development has raised concerns about the sustainability of high-spending AI firms. The startup’s DeepSeek-V3 and DeepSeek-R1, released in January 2024, have sparked an AI race in China, forcing competitors like Alibaba, ByteDance, Baidu, and Tencent to accelerate their own AI innovations.
Following DeepSeek's R1 release, ByteDance updated its flagship AI model, claiming it outperformed OpenAI’s o1 in key benchmarks. DeepSeek, in turn, asserted that R1 matched OpenAI’s model in various performance tests.
DeepSeek’s rise has also triggered an AI price war. When it launched DeepSeek-V2 last year at an unprecedented 1 yuan ($0.14) per million tokens, Alibaba Cloud responded by slashing its own AI pricing by 97%. Other major players, including Baidu and Tencent, followed suit to remain competitive.
Despite the fierce rivalry, DeepSeek’s lean and research-driven approach sets it apart. Founder Liang Wenfeng has emphasized that his startup is not concerned with price battles but is focused on achieving Artificial General Intelligence (AGI)—AI capable of surpassing human abilities in most tasks.
With AI giants facing increasing pressure from agile startups, the landscape of China’s AI industry is evolving rapidly, pushing the boundaries of AI capabilities and global competition.
Comments (0)