Qwen3.8-Flash-Next
- ID
- 18277
- Status
- summarized
- Published
- 26 Aug 2026, 8:52 PM
- Fetched
- 28 Aug 2026, 7:29 PM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://qwen.ai/blog?id=qwen3.8-flash-next
- Source URL
- https://hnrss.org/best
Summary
- Score
- 6.5
- Created
- 28 Aug 2026, 8:35 PM
- Tags
- Audience
- developersai-ml-learnersai-agent-users
What happened
Qwen released open weights for Qwen3.8-Flash-Next, a multimodal MoE model that previews the architecture planned for Qwen4. It introduces four architectural changes: a Gated DeltaNet + Qwen Sparse Attention hybrid that compresses history and uses a lightweight indexer for long-context attention, a Gated Residual design splitting the residual stream into 4 branches, an N-gram Embedding that offloads an embedding table to host memory via async prefetching, and the Muon optimizer refined for orthogonalization accuracy.
Why it matters
If you ship Qwen-family models in production or agents, this preview lets you benchmark the new attention and embedding offload design before Qwen4 lands — the N-gram embedding offload to host memory and sparse attention indexer could materially change your inference cost on long contexts. Builders running local or self-hosted inference should test whether the claimed efficiency gains hold on their hardware.
Discussion angle
Does the N-gram embedding offload to host memory actually help in practice for self-hosted inference setups common in the region, or is it only a win at Alibaba's scale?