Youchuan V8.2 Image-to-Image
Youchuan V8.2 re-imagines an input image guided by a text prompt, returning four variations. Supports native 2K HD, style reference, and aspect-ratio / stylize / chaos / weird controls.
What it does
Youchuan V8.2 Image-to-Image, in practice.
> About Youchuan (悠船): "Youchuan V8.2" is Midjourney's V8.2 model delivered through Youchuan (悠船), Midjourney's officially authorized platform in China — self-branded *"Midjourney 中国实验室" (Midjourney China Lab)* and operated by Xiaochuan Creative (Shanghai) Network Technology Co., Ltd. (小船创意). Youchuan is a legitimate, licensed distributor of Midjourney's models (not a third-party clone): it offers the same Midjourney technology, localized for the Chinese market with priority access to new model versions. Youchuan V8.2 Image-to-Image (youchuan/v8.2/image-to-image) generates four new images guided by an input image together with a text prompt. Midjourney treats the supplied image as a *visual
- Image-guided generation: An input image steers composition, subject, and aesthetic while the text prompt directs the outcome. The image is used as inspiration, not copied exactly — ideal for variations and creative reinterpretation.
- Image prompts restored in V8.2: Image-prompt conditioning (and Midjourney's internal image-weight handling) were absent in the V8.0 alpha and reinstated in V8.2, returning image-driven workflows to the newest model.
- Native 2K HD: With `hd` enabled, V8.2 renders directly at 2048px without a separate upscaling pass.
- ~4–5× faster generation than earlier Midjourney versions (Midjourney-stated), from the GPU-native PyTorch rewrite.
- Optional style reference: A separate style-reference image (`sref`) can be supplied to drive the look (colors, medium, texture, lighting) independently of the content image.
- Aesthetic controls: `stylize`, `chaos`, and `weird` shape how strongly Midjourney's house aesthetic, variety, and unconventionality are applied.
Run Youchuan V8.2 Image-to-Image
from $0.13/imageParameters
What you can set.
| Parameter | What it does | Options |
|---|---|---|
image | Input image to re-imagine. A publicly reachable https image URL (or upload). Each task returns 4 images. | |
prompt | Text prompt guiding the generation. Maximum 1024 characters. Each task returns 4 images. | |
sref | Optional URL of a style-reference image. Must be a publicly reachable https image URL. | |
aspect_ratio | Aspect ratio of the generated image. | 1:1 9:16 16:9 4:3 3:4 2:3 3:2 9:21 21:9 |
hd | Enable native 2K HD generation. Costs 1.5x. | default |
stylize | Controls how strongly the model's default aesthetic is applied (0-1000). | default |
chaos | Higher values produce more unusual and varied results (0-100). | default |
weird | Makes results quirky and unconventional (0-3000). | default |
quality | Controls image detail. 4 = higher detail at the same price (slower). v8.2 supports 1 or 4 only. | 1 4 |
seed | Use the same seed and inputs to get reproducible results. -1 = random. | default -1 |
Sample prompt
The prompt behind the sample.
Convert the image style to a wasteland style.
aspect_ratio: 1:1hd: stylize: chaos: weird: quality: 1seed: -1enable_base64_output: FAQ
Short answers.
How much does Youchuan V8.2 Image-to-Image cost?
Pricing starts at $0.13 per image at the base resolution; higher resolutions and longer durations cost more. Usage is billed per request from your balance — no subscription.
Does Youchuan V8.2 Image-to-Image run uncensored here?
No. Midjourney applies its own content policy to this model regardless of where it is called from. For a route without an extra platform filter use Uncensored MiniMax H3; this page lists Youchuan V8.2 Image-to-Image for everything else it does well.
What does Youchuan V8.2 Image-to-Image take as input?
It is a image editing model. Youchuan V8.2 re-imagines an input image guided by a text prompt, returning four variations. Supports native 2K HD, style reference, and aspect-ratio / stylize / chaos / weird controls.
How do I use it?
Enter an invite code to open Chat with the model loaded, or join the waitlist. We open seats in batches and track which models are requested most.
Related