On February 25, building on the momentum of open-sourcing the Qwen3.5-397B-A17B model on the eve of the Chinese New Year, Ali took another leap forward by open-sourcing three cutting-edge models within the Qwen 3.5 series: Qwen3.5-35B-A3B, Qwen3.5-122B-A10B, and Qwen3.5-27B. These medium-sized models represent a significant leap in performance, outshining their larger predecessors, the Qwen3-235B-A22B and Qwen3-VL, and delivering markedly superior results compared to GPT-5 mini across multiple benchmarks. Presently, the Qwen3.5-Flash model, which is built upon the foundation of Qwen3.5-35B-A3B, is accessible on Alibaba Cloud's BaiLian platform, boasting input prices as competitive as 0.2 yuan per million Tokens.
