Qwen3.8 Max

Qwen3.8-Max is the flagship release from Alibaba's Qwen team, positioned as the most capable model in the family to date. It builds on the architectural foundation of Qwen 3.5 and scales to 2.4 trillion total parameters with 95 billion active, a sparse design that lets the model carry broad general capability while keeping a relatively small active compute footprint per token. The model is described as a multimodal foundation that natively supports visual intelligence alongside text, and it ranks prominently in both Text Arena and Vision Arena. Notably, this is the first Qwen-Max-class model whose weights are planned to be open-sourced, with the open release following shortly after the initial rollout.

In practical terms, Qwen3.8-Max is aimed at end-to-end delivery of complex, multi-step work rather than single-turn question answering. The Qwen team highlights comprehensive improvements across coding, real-world work, research, and long-horizon tasks, with a framing of a model that can take a project from an empty folder to a finished result by writing and running code over extended sessions. The release also emphasizes dependable output for extended workflows, and the same model powers QwenWork, Alibaba's all-in-one workplace AI agent platform. Together, these signals point to a fit for developers and teams looking for a single foundation that can handle coding agents, sustained analytical work, and visually grounded understanding in one integrated system.