WAN 2.7 image edit: text-guided edits with 1–4 reference images, optional prompt expansion, bilingual instructions, and preset output sizes.
Black Forest Labs' FLUX 3 Image Klein revises one to ten pictures at 768, 1K, 1.5K, 2K, or 4K, at the same rate as a new still.
Trading a mask upload for a written instruction, FLUX 3 Image Klein helps product, catalog, and campaign teams change a photo without rebuilding it.
For developers, FLUX 3 Image Klein on RunComfy can be used both in the browser and via an HTTP API, so you don't need to host or scale the model yourself.
Ideal for: Product Photo Edits | Style Transfers | Region Notes
Black Forest Labs built FLUX 3 Image Klein as the open-weight still model in the FLUX 3 family, at 5.6 billion parameters. One instruction plus one to ten pictures becomes a single image. Five size tiers are on this page: 768, 1K, 1.5K, 2K, and 4K. A new still and an edit share those rates.
This page starts from pictures. Leave the aspect on auto and FLUX 3 Image Klein follows the first upload. Pick another ratio when the frame should change. A short box list can sit at the end of the prompt when one region should move or stay. There is no mask file.
The text page is the other door. Open it when FLUX 3 Image Klein should invent the frame from words alone. Stay here when the job already has photographs, product shots, or a style reference.
These controls belong to the edit page. The text page drops the image list and defaults the aspect to 1:1.
| Parameter | Required | Type | Default | Range / Options | Description |
|---|---|---|---|---|---|
images* | Yes (*) | array | Example still | 1-10 image URLs | Source pictures, in the order the instruction names them. |
prompt* | Yes (*) | string | Example line | Text | What should change, and what should stay. |
aspect_ratio | No | string | auto | auto plus the listed ratios, including 9:21 | Output frame. auto follows the first image. |
resolution | No | string | 1k | 768sq, 1k, 1.5k, 2k, 4k | Size tier. This is the only setting that changes the price. |
output_format | No | string | jpeg | jpeg, png | File type of the returned still. |
steps | No | integer | 16 | 1-16 | Sampling steps. The price does not change. |
safety_tolerance | No | integer | 2 | 0-4 | 0 is the strictest filter. |
seed | No | integer | Empty | 0-4294967295 | Repeat a result when the other settings stay the same. |
FLUX 3 Image Klein bills one output image. Reference pictures, aspect ratio, steps, seed, file type, and safety tolerance add no separate fee. The text page uses the same table.
| Resolution | Price per image |
|---|---|
| 768 | $0.007 |
| 1K | $0.008 |
| 1.5K | $0.009 |
| 2K | $0.013 |
| 4K | $0.035 |
A 1K edit is the usual first charge on FLUX 3 Image Klein.
Based on publicly available information. These notes describe inputs, not a scored ranking.
| Model | Inputs | Size | Standout |
|---|---|---|---|
| FLUX 3 Image Klein | One to ten pictures on this page, or text alone on the text page | 768, 1K, 1.5K, 2K, 4K | Same rate for a new still and an edit, optional region note in the prompt |
| FLUX 3 Image | Text, or an edit with up to ten references on its own pages | 1K, 2K, 4K | The larger commercial still model, without the 768 and 1.5K tiers |
| Earlier FLUX still models | A shorter prompt and fewer size choices | Smaller published frames | Useful when a workflow is already built on them |
Stay on this page when the source is a picture. Open the text page when FLUX 3 Image Klein should start from words.
WAN 2.7 image edit: text-guided edits with 1–4 reference images, optional prompt expansion, bilingual instructions, and preset output sizes.
WAN 2.7 Pro image edit: high-fidelity prompt-driven edits with 1–4 references, prompt expansion, and the same controls as the standard edit endpoint.
Edit up to ten images from a written instruction at 1K, 2K, or 4K.
FLUX 3 Image Klein revises one to ten pictures from a written instruction. Typical work is a product photo edit, a style transfer, or a small region change that should leave the rest of the frame alone.
FLUX 3 Image Klein accepts one to ten image URLs on this page. The first upload is image 1. Refer to them in that order inside the instruction. Each request still returns one image.
No. FLUX 3 Image Klein does not take a mask file. When a region matters, describe it in the instruction, or add a box note on the 0 to 1000 grid inside the same prompt.
The edit uses 768, 1K, 1.5K, 2K, and 4K. auto follows the first image. The other choices match the text page, including wide and tall ratios through 9:21. The returned pixel size depends on the ratio you pick.
Say what must stay and what must change. FLUX 3 Image Klein follows that split more reliably when the instruction names the uploads in order and keeps the request to one change.
No. FLUX 3 Image Klein bills the output tier only. One reference or ten costs the same. Steps, seed, aspect ratio, and file type also leave the price unchanged.
Yes. Try the edit in the RunComfy model UI, then send the same images, instruction, and settings to the HTTP API. The API uses those parameters. It does not promise a fixed speed.
An edit costs the same as a new still: $0.007 at 768, $0.008 at 1K, $0.009 at 1.5K, $0.013 at 2K, and $0.035 at 4K. Runs spend usd or credits from the account balance.
RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.





