Alibaba’s Qwen family continues to impress with Qwen 3.7, the latest in a series of competitive open-weight models. Qwen 3.7 builds on the success of Qwen 3 with improved reasoning, longer context, and better multilingual performance.
Model Family
Qwen 3.7 comes in several sizes: the 72B balanced model at $0.6/$2 per million tokens, the larger Max variant for maximum capability, and the Flash variant for speed. All are open-weight under a permissive license.
Multilingual Strength
Qwen’s strongest differentiator is multilingual performance. It excels at Chinese, Japanese, Korean, and other Asian languages while maintaining strong English capabilities. For organizations serving Asian markets, Qwen is often the best choice among open models.
Best For
Asian language applications, cost-sensitive deployments, and teams that want open-weight flexibility. Qwen 3.7 is particularly strong for Chinese-language content generation, translation, and analysis while remaining competitive for general English tasks.
Related: AI Model Comparison 2026
