AIMarkTechPostabout 1 hour ago
Alibaba Qwen Releases Qwen3.8-Flash-Next Model
4 min read
Alibaba's Qwen team released Qwen3.8-Flash-Next, an open-weight multimodal Mixture-of-Experts model with a 125B backbone and 6B active parameters per token. It previews the Qwen4 architecture, featuring a Gated DeltaNet and Qwen Sparse Attention hybrid, Gated Residual, N-gram Embedding, and the Muon optimizer.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
