Prompt History
0 saved
💡 Tip: Click any <imageN> tag below to insert it at the cursor position.
Input Images
0 / 10
Upload or drag images here, or paste from clipboard (Ctrl+V)
Up to 10 images (PNG, JPG, WEBP). Refer via <image1> to <image10>
1024 × 1024
40 steps
#
Advanced API Settings â–¼
Target Mode: /v1/images/generations
Inference Time:
-
Applied Seed:
-
Output Size:
-
Mode:
-
Generated Output
Ready to synthesize
Enter a prompt and click "Generate Image". With input images attached, Qwen-Image-2.1 performs multi-image instruction editing.
Generating with Qwen-Image-2.1...
Elapsed: 0.0s
Text encoding & DiT denoising (~40-50s on RTX 3090 with CPU offload)
Generation Failed
Qwen3-VL 8B + 7B Single-Stream DiT (32 layers)
RGBA & Multi-Image Editing Supported