Alibaba Releases Qwen3.8-Flash-Next, a 125B MoE Preview of Qwen4 That Activates 6B Parameters Per Token

Alibaba's Qwen team has opened the weights of Qwen3.8-Flash-Next, a multimodal mixture-of-experts model with 125 billion parameters that activates only 6 billion per token, as an experimental preview of the architecture planned for Qwen4. Qwen reports that the checkpoint outperforms the much larger Qwen3.7-Plus on coding and office evaluations while training at about one-ninth the cost.