Alibaba is set to release Qwen 3.8-Flash-Next on Wednesday, an artificial intelligence model with 125 billion total parameters that activates only 6 billion per token. The Qwen team describes this model as a preview of the Qwen 4 architecture, the next generation of their model family. The system uses a mixture-of-experts design, which activates only the specialized sub-models needed for each task, reducing computational requirements. Official benchmark scores have not yet been published and the model weights were not available on ModelScope at the time of writing. This release follows the rapid pace of open-weight models from Chinese developers, alongside DeepSeek and Moonshot.
Source: Read the original article
