Alibaba unveils latest Qwen model to boost global AI adoption

Alibaba just dropped not one, not two, but three new AI models in quick succession, and the numbers attached to them are staggering. The flagship Qwen3.8-Max contains 2.4 trillion total parameters, making it one of the most powerful publicly available models on the planet. What Alibaba actually shipped The rollout came in three waves. Qwen3.8-Max arrived first around August 3, built on a mixture-of-experts (MoE) architecture. That means while the model has 2.4 trillion parameters in total, only 95 billion are active at any given time. The model has already earned a fifth-place ranking in Text Arena and second place in Vision Arena, with particularly strong performance on coding tasks. Next came Qwen3.8-27B, released between August 14-17. This is the lightweight sibling, sporting a comparatively modest 27 billion parameters. It’s designed to run on consumer hardware, including laptops. The model recorded over 1 million downloads shortly after launch. The final piece landed on August 26 with Qwen3.8-Flash, a 125 billion parameter model positioned as the budget-friendly option. Alibaba is pitching it directly against Anthropic’s Opus and DeepSeek’s V4-Flash. The open-weight playbook A...