Search
In September 2025, the Ali Tongyi Qwen team unveiled its most substantial and potent model to date: the Qwen3-Max. This model is presented in three distinct versions—Base, Instruct, and Thinking. Among these, the officially launched Instruct version shines in code generation and agent tasks. Meanwhile, the Thinking version, which is still undergoing training, has already showcased remarkable reasoning abilities. At present, the official Qwen3-Max-Instruct version can be accessed on Qwen Chat. It provides API access and supports more than 100 languages. This model excels in mathematical reasoning, programming, and scientific endeavors. It is capable of comprehending and executing complex instructions with greater accuracy, minimizing knowledge hallucinations (instances where the model generates incorrect or made-up information), and generating responses of superior quality.
-
C114 Communication Network -
Communication Home
