Chinese tech giant Alibaba Group Holdings Ltd.’s Qwen model is emerging as a cost-effective alternative to DeepSeek, as U.S. computer scientists successfully develop a new reasoning model for under $50 using Alibaba’s open-source technology.
Researchers from Stanford University and the University of California, Berkeley leveraged Alibaba’s Qwen 2.5-32B-Instruct model to reduce AI training costs, according to a recent research paper.
Qwen is an open-source AI model developed by Alibaba Group, designed for cost-effective and high-performance AI applications. The Qwen 2.5 series, launched in September, includes models ranging from 0.5 billion to 72 billion parameters, with strong capabilities in reasoning, mathematics, and coding. Supporting over 29 languages.
Qwen offers a scalable alternative to other AI models like DeepSeek R1. Its affordability has been demonstrated by U.S. researchers, who successfully trained an AI model for just $50 using Qwen, highlighting its potential to make AI development more accessible worldwide.
Experts highlight Alibaba’s AI advancements as further proof of China’s growing influence in the rapidly evolving artificial intelligence sector. This development follows the surprise success of Chinese AI startup DeepSeek, which recently launched its high-performance, cost-efficient open-source AI model R1.
