26 August 2026

Qwen releases early preview of next-generation multimodal model

• Qwen, a Chinese AI lab, released Qwen3.8-Flash-Next, a model that can process both text and images with open weights (publicly available code).
• The model uses a mixture-of-experts architecture, meaning it activates different specialized components for different tasks, reducing computational cost while maintaining performance.
• Qwen positioned the release as an early look at the architecture planned for Qwen4, its next major version.

How it was covered

Simon WillisonDaily notes and links

Qwen released a new open weights multimodal mixture-of-experts model with 125B total tokens but only 6B active parameters, providing significant performance improvements. The model serves as an early preview of Qwen4's architecture and the newsletter author has been testing quantized versions on a DGX Spark.