Qwen

Qwen is Alibaba's family of open-source large language models, positioned alongside DeepSeek as one of China's leading open-source AI offerings . The model series has been adopted across the Chinese AI ecosystem: SiliconCloud's platform uses Qwen2 for text dialogue alongside other mainstream models , and Peak's Steiner reasoning model was built on Qwen2.5-32B and Qwen2.5-32B-Instruct bases as of late 2024 . DeepSeek also distilled its R1 model into Qwen variants ranging from 1.5B to 32B parameters, a move described by one commentator as demonstrating Qwen's substantial influence within China . By late 2025, Zhang Tao of Manus cited Qwen and DeepSeek as evidence that open-source models had closed ground on OpenAI, enabling a more flexible environment for application builders .