FREE starter on Whop (lite workflow + exclusive cheat sheet): https://whop.com/imagelab/free-starter-qwen-image-2512-edit-2511-workflows-pack/
FULL pack ($39): https://whop.com/imagelab/qwen-image-2512-edit-2511-workflows-pack/ — use code QWN230 for 30% off ($39 → $27.30)
Alibaba's Dec-2025 Qwen-Image update: photoreal people, natural detail, and accurate text layout in images. Apache 2.0.
## What you get free
Qwen2512_Lite_T2I.json — 4-step Lightning text-to-image at 1:1 (1328×1328). Complete and working — no locked nodes, no watermark.
## FREE vs FULL
| | FREE (here) | FULL pack ($39) |
|---|---|---|
| Workflows | 1 (Lightning 4-step T2I, 1:1) | 5 (50-step T2I + Lightning + Edit 2511 40-step + Edit material + ControlNet Union) |
| Edit 2511 (instruction editing) | — | ✅ |
| ControlNet Union layout control | — | ✅ |
| Aspect ratios | 1:1 only | All 7 (1:1 to 16:9) |
| Prompt library | — | ✅ |
| Guides | This description | Full EN + AR guides |
## Setup
1. ComfyUI (recent version — standard nodes, no custom pack needed).
2. Model files: qwen_image_2512_fp8_e4m3fn.safetensors (diffusion_models), qwen_2.5_vl_7b_fp8_scaled.safetensors (text_encoders), qwen_image_vae.safetensors (vae), Qwen-Image-Lightning-4steps-V1.0.safetensors (loras). From HuggingFace (Qwen).
## License
Apache 2.0, except the text encoder under the Tencent community licence. Models ~31 GB; fp8 ~24 GB VRAM.

