Qwen-Image-Layered turns a flat image into separate transparent layers, giving an artist room to revise props, adjust a composition or make variants. Its appeal is strongest when hand-cutting those elements would take too long. The cleanup still matters: check overlaps, reconstructed hidden areas, edge colors and alignment when the layers are reassembled. Treat the result as source material for compositing and manual finishing, with no guarantee of a usable animation rig or semantic scene graph. Layered uses a distinct model and workflow, so check compatibility before bringing in a Qwen generation or editing LoRA.
Commercial use & license
The official model card lists Apache 2.0. Rights to the input artwork and reuse of the resulting layers remain separate.
Adult content rules
Explicit-content eligibility remains unverified. Local execution leaves artwork rights unresolved; hosted use also requires checking the provider’s terms.
What you need
A compatible GPU runtime, the model components and enough storage; measure memory use in your chosen workflow.
Ranks are recommendations for the task described, based on reviewed documentation. Model ranks cover non-explicit development work. We have not run practical benchmarks. Check content policies and commercial terms separately. How the ranks work →