Text to image and image editing in one model
Leave the reference images empty to generate from text, or add 1 to 10 images to edit, combine, or restyle them with a written instruction.
Use Seedream 5.0 Flash on AIReiter: fast text-to-image, editing with up to 10 references, and image-to-layers splitting at 1K to 2K, online or via API.
| Specification | Credits | Price (USD) |
|---|---|---|
| Model & ModalitySeedream 5.0 Flash1K / 1.5K / 2K | 2.5per image | $0.025per image |
| Model & ModalitySeedream 5.0 Flash LayerizeLayerize (base image + each layer) | 2.5per image | $0.025per image |
Prices follow the model configuration. Reference images are not billed unless the model says so.
Seedream 5.0 Flash is ByteDance's fast, low-cost model in the Seedream 5.0 family. Generate from text, edit with up to 10 reference images, or split a finished image into editable layers, all on one page and through one API.
Leave the reference images empty to generate from text, or add 1 to 10 images to edit, combine, or restyle them with a written instruction.
Pick the resolution you need without paying more for a sharper image. Reference images are not billed.
Draw boxes, arrows, or coordinates on a reference image to show exactly where a change should happen.
Switch to the Layerize tab to split one image into a base image plus up to 16 transparent PNG layers, each with a name, description, and position.
Switch between the tabs at the top of the page. Each mode has its own form and its own API model ID.
Text-to-image and image-to-image. Prompt, 0 to 10 reference images, 8 aspect ratios, 1K / 1.5K / 2K, JPEG or PNG. Billed per image when you submit.
Upload one PNG or JPEG image and optionally name the elements to separate. Returns a base image and up to 16 layers. Billed per output image after the result is ready.
Use it where speed and volume matter as much as image quality.
Produce banners, ads, thumbnails, and posts in many variations quickly, in 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 1:1, or 21:9.
Keep a product consistent across scenes by passing it as a reference image, then change the background, lighting, or props in words.
Render readable text in many languages, including Japanese, Korean, French, German, Spanish, Portuguese, and Arabic, for posters and ads per market.
Turn a flat poster or illustration into separate layers you can move, resize, or replace in a design tool.
The playground and the API expose the same parameters.
Up to 10 JPEG, PNG, or WebP images, up to 30 MB each. Layerize takes exactly one PNG or JPEG image.
1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, or 21:9. The default is 1:1. Layerize keeps the aspect ratio of the source image.
1K, 1.5K, or 2K, all at the same price. The default is 2K. Layerize also offers Auto, which follows the source image size.
JPEG or PNG. In Layerize this sets the base image format; the separated layers are always transparent PNG.
Both modes are billed per output image. See the pricing table on this page for current credits.
Text-to-image and image-to-image cost the same at 1K, 1.5K, and 2K. Reference images are free.
The base image and every separated layer each count as one image, up to 17 per request. You are charged after the result is ready, based on the actual number of images.
A Layerize task needs enough balance to cover the largest possible result before it starts. Only the actual output is charged.
If a generation fails, no credits are charged.
Submit one normalized task and query the same external task ID until it reaches a terminal state.
Send seedream_v5_flash to generate or edit images, or seedream_v5_flash_layerize to split an image into layers.
POST /api/openapi/submit with the model, a unique out_task_id, and params: prompt, aspect_ratio, resolution, output_format, and optionally image_url (up to 10 for generation, exactly 1 for Layerize).
POST /api/openapi/query with the same out_task_id until the status is completed or failed. Layerize results list the base image and every layer with its z-index, name, description, and bounding box.
Download the output images and keep them in your own storage instead of hot-linking the result URLs long term.
Compare Seedream 5.0 Flash with other image models on AIReiter.
The quality-first model in the Seedream 5.0 family. Choose Pro when fine detail matters more than speed and cost.
Layer decomposition built on Seedream 5.0 Pro, for when layer quality matters most.
A fast image model with accurate in-image text and up to 10 reference images, with output up to 4K.
| Control | Type | Default | Notes | |
|---|---|---|---|---|
| 01 | prompt | text | — | — |
| 02 | image_url | image | — | — |
| 03 | aspect_ratio | string | 1:1 | — |
| 04 | resolution | string | 2K | — |
| 05 | output_format | string | jpeg | — |
Use Seedream 5.0 Flash to generate or edit images, or Seedream 5.0 Flash Layerize to split an image into layers.
Describe what you want. Add reference images to edit or combine them, or upload the image to split in Layerize.
Choose the aspect ratio, resolution, and output format.
Download the image, or in Layerize toggle and inspect each layer, then download one layer or all of them.
seedream_v5_flash, seedream_v5_flash_layerize
POST to /api/openapi/submit with the model, a unique out_task_id, and parameters.
POST to /api/openapi/query with the same out_task_id until the task completes.
Save outputs in your own storage rather than hot-linking them.
Seedream 5.0 Flash is ByteDance's fast, cost-efficient image model in the Seedream 5.0 family. It supports text-to-image, image editing with up to 10 reference images, and layer decomposition.
They share the same core features. Flash focuses on speed and lower cost for high-volume work, while Pro focuses on the highest image quality.
Use the same Seedream 5.0 Flash model. Leave the reference images empty for text-to-image, or add 1 to 10 images for image-to-image.
1K, 1.5K, and 2K, at the same price. Aspect ratios are 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, and 21:9.
A base image plus up to 16 transparent PNG layers. Each layer has a name, a description, and a bounding box showing where it sits in the original image.
Per output image after the result is ready. The base image and each layer count as one image, up to 17 per request, so the cost depends on how many layers the model finds.
No. Only output images are billed, and failed tasks are not charged.
Use seedream_v5_flash for generation and editing, and seedream_v5_flash_layerize for layer decomposition, with POST /api/openapi/submit and then POST /api/openapi/query.