In a surprise move on the first day of the Lunar New Year, Chinese tech giant Alibaba (9988.HK) launched its latest artificial intelligence model, Qwen 2.5-Max, claiming it has outperformed rivals including DeepSeek’s highly praised DeepSeek-V3. The launch, announced by Alibaba’s cloud unit on its official WeChat account, comes as the AI landscape in China is rapidly evolving, with DeepSeek’s recent success sparking a scramble among domestic competitors to upgrade their models.
Alibaba’s new Qwen 2.5-Max was positioned as a major leap over other leading models, including OpenAI’s GPT-4, DeepSeek-V3, and Meta’s Llama-3.1-405B, with the company highlighting its model’s superior performance across various benchmarks. This release follows the impressive performance of DeepSeek, a Chinese AI startup that has captured global attention with its low-cost and high-performance models.
DeepSeek’s rapid rise, especially after its DeepSeek-V2 model triggered a price war in the AI industry last May, has forced even the biggest players in China to react. Just days after DeepSeek’s launch of its R1 model in January, ByteDance released an update to its own AI, claiming it surpassed OpenAI’s o1 in AIME, a benchmark for AI comprehension.
Liang Wenfeng, DeepSeek’s founder, has expressed his belief that large tech companies like Alibaba may not be the best equipped for the future of AI, citing their higher operational costs and hierarchical structures compared to DeepSeek’s lean, research-driven model. Despite this, Alibaba’s new Qwen model signifies its resolve to maintain a competitive edge in a market increasingly driven by innovation from startups like DeepSeek.
The fierce competition among AI companies in China comes as DeepSeek continues to challenge the dominant players in the sector, raising doubts about the sustainability of massive spending by U.S. firms. Investors are closely watching how these domestic battles will reshape the future of AI development, both in China and globally.