devtools.sh

Qwen3.8 Flash Next

Qwen

An open-weights multimodal MoE model that doubles as an early preview of the Qwen4 architecture, the same role Qwen3-Next played for Qwen3.5. It pairs Gated DeltaNet with Qwen Sparse Attention, widens the residual stream into four gated branches, and adds 51B N-gram embedding parameters that cost almost nothing per token. Natively 262K context, extensible to 1M with YaRN.

Model page

Specs

Model

Family
Qwen
Variant
3.8 Flash Next
Parameters
125B total, 6B active, plus 51B N-gram embeddings
Weights
Open
License
Qwen Community License 1.0

Limits

Context
262,144 tokens

Modalities

Input
Text Image Video
Output
Text

Capabilities

Reasoning
Yes
Reasoning control
On / off, Effort levels
Tool calling
Yes
Attachments
Yes

Dates

Announced
24 Aug 2026
Released
24 Aug 2026

More from Qwen

All models

Qwen3.8 27B

Qwen

Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability.

Apache-2.012d ago

Keep up with the tools

An occasional email when notable AI dev tools and models land in the directory. No spam, unsubscribe anytime.