Open Models Almanac
Methods

general family · not yet filed

Gated DeltaNet and Qwen Sparse Attention hybrid architecture

The sources on this page only mention it in passing.

Also called GDN + QSA hybrid architecture.

sources
2
model
1
lab adopt it
1
strongest
core

Documented in

Alibaba (Qwen)

Evidence

2 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
core 2

Qwen3.8-Flash-Next

Alibaba (Qwen)adopted2 sources

A GDN + QSA hybrid architecture. Gated DeltaNet (GDN) compresses the history efficiently; Qwen Sparse Attention (QSA) uses a compressed lightweight indexer to select the important context at micro-block granularity

coremodel architecturein Qwen3.8-Flash-NextQwen Team
1 more from this model

out of every 4 layers, 3 GDN layers compress history into a fixed-size state, while the remaining layer performs precise retrieval over the full context.

coreunclearin Qwen3.8-Flash-NextQwen Team
↑