- Alibaba launches Qwen 2.5-Max, claiming it surpasses DeepSeek-V3.
- The release coincides with Lunar New Year, signaling urgency from domestic competitors.
- DeepSeek’s rapid rise has prompted a race among Chinese tech giants to upgrade AI models.
- Alibaba’s Qwen 2.5-Max outperforms leading AI systems like GPT-4o and DeepSeek-V3.
- DeepSeek’s low-cost, open-source models have sparked a price war in China’s AI sector.
BEIJING, Jan 29, 2025 – In an unprecedented move, Alibaba’s cloud division has released the Qwen 2.5-Max, claiming it outperforms DeepSeek’s highly-praised DeepSeek-V3 model. The release comes amid fierce competition in the Chinese AI market, which has been disrupted by the rapid rise of DeepSeek, a startup that has generated significant buzz globally.
The timing of Qwen 2.5-Max’s launch—on the first day of the Lunar New Year—adds to its significance. Typically, this is a period when most Chinese workers are off for the holiday, making the decision to unveil a major product even more impactful. It reflects the mounting pressure that DeepSeek’s unexpected success has put on both domestic and international competitors. Alibaba’s cloud unit, in a statement posted on WeChat, declared that the new model outperforms major AI systems, including OpenAI’s GPT-4o, DeepSeek-V3, and Meta’s Llama-3.1-405B.
DeepSeek’s Impact on the AI Landscape
DeepSeek’s rapid rise began with the release of its DeepSeek-V3 model on January 10, 2025. Within just weeks, the Chinese startup became a formidable player in the AI race, not just within China but globally. The launch of DeepSeek’s AI assistant, powered by V3, shocked the tech community. Industry experts and investors alike took note of the model’s performance and its unusually low development costs. DeepSeek’s models have offered an alternative to more expensive AI systems, prompting investors to reassess the heavy spending plans of larger companies like OpenAI.
On January 20, just days after the release of DeepSeek’s R1 model, TikTok parent ByteDance quickly followed suit, unveiling an update to its AI model. ByteDance claimed its updated model outperformed OpenAI’s GPT-4 on AIME, an important benchmark for assessing AI’s ability to understand and respond to complex instructions. This update echoes DeepSeek’s earlier claims about the competitiveness of its R1 model.
The Price War: How DeepSeek Changed the AI Market
DeepSeek has had a big impact on the AI market, especially with its pricing strategy. In May 2024, the company released its V2 model at an open-source price of just 1 yuan ($0.14) per million tokens. This surprised the industry and forced competitors like Alibaba, Baidu, and Tencent to lower their prices sharply. Alibaba’s cloud division, for example, reduced its model prices by up to 97%.
Liang Wenfeng, the founder of DeepSeek, downplays the focus on price wars. In an interview with Chinese media outlet Waves in July 2024, he said DeepSeek’s goal isn’t to win by price. Instead, the company aims to push forward the development of Artificial General Intelligence (AGI), which he believes is the true challenge.
AGI refers to autonomous systems that can outperform humans in most important tasks. Many AI experts see AGI as the future of technology. Liang’s statement reflects his commitment to achieving this milestone, not just competing on cost.
DeepSeek’s Lean Operations vs. Tech Giants
DeepSeek operates with a lean, research-driven model, staffed mainly by young graduates and PhD students from top Chinese universities. In contrast, the large tech companies like Alibaba employ hundreds of thousands of workers and follow top-down management structures. Liang has argued that this makes the traditional tech giants ill-suited for the future of AI. He stated that DeepSeek’s agility and focus on innovation give it an edge over larger firms, which are constrained by their size and established practices.
Liang’s strategy differs significantly from that of larger companies. DeepSeek has focused on continued innovation and flexibility, prioritizing breakthroughs in AGI over massive corporate structures and budgets. According to Liang, “Large foundational models require continual innovation, but the capabilities of large tech companies have limits.” This approach has garnered DeepSeek a loyal following, despite the challenges posed by its larger, more established competitors.
Alibaba’s Response and Future Strategy
In response to DeepSeek’s pressure, Alibaba has launched the Qwen 2.5-Max, claiming it outperforms DeepSeek-V3 and other AI models. Alibaba’s cloud division praises the new model’s capabilities. However, it remains to be seen if this claim is true, as DeepSeek continues to rapidly improve its own models.
The competition in China’s AI sector is likely to speed up innovation. As companies release new updates and improvements, the focus may shift to more efficient, cost-effective solutions. DeepSeek’s innovative pricing strategy forces large tech companies to rethink their traditional business models.
For Alibaba and its competitors, staying ahead is about more than technology. It’s also about keeping market share in a fast-changing landscape. The battle for AI supremacy will shape not only China’s tech future but also the global technology scene.