SOURCE-LINKED INTELLIGENCE
Qwen: Qwen3.5 397B A17B
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...
Read original source ↗ Open in workspace
- recordType
- model-catalog
- evidenceStatus
- provider-reported
- region
- Global
Evidence & attribution
First collected: 2026-09-20T08:01:03.945Z. This is not the publication date.