On February 25, 2026, Alibaba made a significant announcement regarding the expansion of its Qwen3.5 open-source family. Building on the initial release of the flagship model, Qwen3.5-397B-A17B, the company has introduced a range of new multi-scale versions, namely Qwen3.5-122B-A10B, Qwen3.5-35B-A3B, and Qwen3.5-27B (Dense). Among these additions, Qwen3.5-122B-A10B and Qwen3.5-27B (Dense) have notably bridged the performance gap between small-to-medium-sized models and leading closed-source alternatives, demonstrating exceptional proficiency in complex Agent tasks. The Qwen3.5-35B-A3B model, on the other hand, represents a comprehensive upgrade over its predecessor, while the Qwen3.5-27B (Dense) model achieves remarkable energy efficiency within a compact framework, thereby reducing the entry barriers for large-scale model utilization. Additionally, the Qwen3.5-Flash API is now accessible via Alibaba Cloud's BaiLian platform, offering support for million-token contexts and native toolchains. Researchers and developers can download these models from GitHub, Hugging Face, and the ModelScope community for further research and customization.