Qwen Image LoRA – Composite
Composite model by Qwen — China.
LoRA-powered compositing variant for consistent style transfer across mixed references.
LoRA-guided composites that keep subjects and style consistent while following precise prompts.
Try Qwen Image LoRA – Composite on LORY
Sign in to start generating with this model in your own story project.
Qwen Image Edit LoRA
LoRA-capable edit model that is strong at stylized changes and localized revisions.
Seedream 5 Lite – Composite
Bytedance Seedream 5 Lite multi-reference composition — accepts up to 10 reference images and blends them via prompt-led restyling, with Figure 1 /…
Flux 2 Pro – Composite
Reference-aware image compositing model for combining subjects into coherent final scenes.
Z-Image Turbo – Composite
Speed-oriented compositing model for quick prompt-plus-reference blends.
OpenAI gpt-image-2 – Composite
gpt-image-2 composite profile that blends up to 10 high-fidelity references into a single cohesive scene.
Nano Banana 2 (Flash) – Composite
Gemini 3.1 Flash Image composite profile for fast, high-throughput reference merges.