Chinese tech giant Alibaba has launched Qwen 2.5, an upgraded AI model it claims outperforms top open-source models, including DeepSeek-V3.

“Qwen 2.5-Max outperforms … almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,” Alibaba’s cloud unit stated on WeChat. The release coincided with Lunar New Year celebrations.

The announcement comes amid growing competition in China’s AI market, driven by DeepSeek’s aggressive pricing. Its DeepSeek-V2 model, launched in May 2024, forced major players like ByteDance, Tencent, Baidu, and Alibaba to lower prices.

“The Jan. 10 release of DeepSeek’s AI assistant … as well as the Jan. 20 release of its R1 model, has shocked Silicon Valley … prompting investors to question huge spending plans by leading AI firms in the United States,” an analyst told Reuters.

DeepSeek, founded in May 2023 by Liang Wenfeng, employs an innovative AI training approach and cost-efficient Mixture-of-Experts (MoE) architecture. Its low-cost API pricing—$0.55 per million input tokens compared to OpenAI’s $15—has disrupted the market.

Despite its affordability, DeepSeek lacks advanced AI-generated media tools found in rival models. However, its efficiency raises questions about the sustainability of U.S. tech giants’ massive AI investments.

Vanguard News