Term
Qwen 3.8-Max
Qwen 3.8-Max (August 2026) is Alibabas largest flagship — an MoE model with 2.4T parameters, a 1M-token context and multimodal input, later released with open weights.
Qwen 3.8-Max — explained in more detail
Alibaba previewed Qwen 3.8-Max on July 19, 2026 and officially released it on August 3, 2026 with published specs and API pricing. It is the largest and most capable flagship of the Qwen range to date: a Mixture-of-Experts (MoE) model with 2.4T parameters that processes text, images and video, offers a context window of up to 1M tokens and generates up to 128K tokens per response.
Example / Practical use
In public comparisons, Qwen 3.8-Max ranked fifth in Text Arena and second in Vision Arena, and shows strengths in coding, real-life work, research and long-horizon tasks. The model first became available through QwenCloud and Alibaba Cloud Model Studio; an open-weights release followed a week later. Notably, it is the first time Qwen open-sources a Max-class model — alongside a smaller Qwen3.8-27B. This means the model can be used both through the cloud API and self-hosted.
Distinction
Max is Qwens top-tier line above the leaner Flash and mid-size models (such as Qwen3.8-Flash-Next or Qwen3.8-27B). While Flash targets cost efficiency, Max stands for maximum capability. As a model from China, Qwen competes against the Western flagships from OpenAI, Anthropic and Google, but sets itself apart through the planned release of open weights.