Skip to content
Model infoModelsSnapshot: 15 Oct 2026

ByteDance / Seedream

Seedream 4.0 (250828)

Seedream 4.0 image-generation revision 250828, supporting text and reference-image generation.

Image generation
availableClosed weightsImage generationImage editing
Official documentation (opens in a new tab)Reviewed 15 Oct 2026
10 of 23 specification fields collected
Output resolution1K, 2K, 4KSee specification notes below
Output formatsjpegSee specification notes below
Input billingSee ratesBytePlus ModelArk
Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected
Maximum output
Not collected
Input modalities
Text, ImageSource for Input modalities (opens in a new tab)
Output modalities
ImageSource for Output modalities (opens in a new tab)
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Streaming
SupportedSource for Streaming (opens in a new tab)
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Parameters
Not collected
Architecture
Not collected
Languages
Not collected
Context caching
Not collected
Layer decomposition
Not supportedSource for Layer decomposition (opens in a new tab)
Interactive image editing
Not supportedSource for Interactive image editing (opens in a new tab)
Multiple-image generation
SupportedSource for Multiple-image generation (opens in a new tab)
Output resolutions
1K, 2K, 4KSource for Output resolutions (opens in a new tab)
Output formats
jpegSource for Output formats (opens in a new tab)
Prompt optimization modes
standard mode, fast modeSource for Prompt optimization modes (opens in a new tab)
Published request quota
500 images/minuteSource for Published request quota (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

BytePlus ModelArk · standard

Provider source (opens in a new tab)
API model ID
seedream-4-0-250828
Access
ModelArk API access requires an eligible BytePlus account and enabled model. Published quotas are account-dependent.
Regions
Not collected
Service tier
standardSource for Service tier (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 12 Oct 2026
CategoryUsageRateUnit
Image generationOutput1$0.03image
Image generationInput2$0.00image

Rate conditions

  1. Per successful output image. Multiple-image requests are charged per image, not per request.
  2. No input-image charge in the reviewed price schedule. Output is charged separately.
Pricing source (opens in a new tab)

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

fal

Provider source (opens in a new tab)
API model ID
fal-ai/bytedance/seedream/v3/text-to-image

Deprecated v3 name routed to Seedream 4.0. Native revision is not exposed.

Access
Deprecated compatibility route. Documentation states requests execute Seedream 4.0 and results are resized to the requested dimensions when v4 minimum area requires a larger image. Account and API key required; no inference executed.
Regions
Not collected
Declared backend
Seedream 4.0; exact native checkpoint not statedSource for Declared backend (opens in a new tab)
Endpoint availability
Deprecated v3 selector; transparent v4 replacementSource for Endpoint availability (opens in a new tab)
Image input limits
Documented request width and height: 512–2048 pixels eachSource for Image input limits (opens in a new tab)

Endpoint-specific image_size description takes precedence over the shared ImageSize type’s generic 14142 maximum. Resizing is part of the compatibility behavior.

Output count choices
1–4 images per request; default 1Source for Output count choices (opens in a new tab)
Guidance scale
1–10; default 2.5Source for Guidance scale (opens in a new tab)
Seed control
SupportedSource for Seed control (opens in a new tab)
Image size presets
square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9Source for Image size presets (opens in a new tab)
Configuration discrepancy
Shared ImageSize type allows sides up to 14142, but endpoint description limits sides to 512–2048.Source for Configuration discrepancy (opens in a new tab)

Do not promote the generic schema ceiling into the endpoint’s supported image dimensions.

Generation mode
Queued API with optional synchronous data-URI outputSource for Generation mode (opens in a new tab)

sync_mode omits output data from request history; this does not establish zero retention or no-training terms.

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 15 Oct 2026
CategoryUsageRateUnit
Hosted compatibility image generationOutput image via v3-compatible route1$0.03image

Rate conditions

  1. Published per-image price for the deprecated v3 selector. API documentation says Seedream 4.0 now serves this route, including compatibility resizing. It is not a current native Seedream 3.0 rate.
Pricing source (opens in a new tab)

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Volcengine Ark images

Provider source (opens in a new tab)
API model ID
doubao-seedream-4-0-250828
Access
Deprecated native version. Existing endpoints continue until the scheduled November 24, 2026 shutdown at 14:00 UTC+8; migration behavior differs by version.
Regions
cn-beijing
API endpoint URL
POST https://ark.cn-beijing.volces.com/api/v3/images/generationsSource for API endpoint URL (opens in a new tab)
Input modalities
Text, ImageSource for Input modalities (opens in a new tab)
Output modalities
ImageSource for Output modalities (opens in a new tab)
Reference images
14Source for Reference images (opens in a new tab)
Image input limits
Each reference: at most 30 MB; 196–36,000,000 pixels total; each side above 14 px; aspect ratio 1:16–16:1.Source for Image input limits (opens in a new tab)
Input image formats
jpeg, png, webp, bmp, tiff, gif, heic, heifSource for Input image formats (opens in a new tab)
Image input methods
Accessible image URL or data:image/<lowercase-format>;base64 payload.Source for Image input methods (opens in a new tab)
Output resolutions
1K, 2K, 4KSource for Output resolutions (opens in a new tab)
Resolution controls
Custom width × height: 921,600–16,777,216 total pixels; aspect ratio 1:16–16:1.Source for Resolution controls (opens in a new tab)

These are pixel-product bounds, not independent maximum dimensions.

Output formats
jpegSource for Output formats (opens in a new tab)
Output link expiry
Returned image URLs expire after 24 hours; response_format can instead return b64_json.Source for Output link expiry (opens in a new tab)
Streaming
SupportedSource for Streaming (opens in a new tab)

For supported image models, stream=true returns each completed image; not token-by-token chat output.

Multiple-image generation
SupportedSource for Multiple-image generation (opens in a new tab)
Generation mode
sequential_image_generation=auto enables related image sets; input references plus outputs must total at most 15.Source for Generation mode (opens in a new tab)
Layer decomposition
Not supportedSource for Layer decomposition (opens in a new tab)
Interactive image editing
Not supportedSource for Interactive image editing (opens in a new tab)
Prompt optimization modes
standard, fastSource for Prompt optimization modes (opens in a new tab)
Published request quota
500 generated images/minute per account and model version.Source for Published request quota (opens in a new tab)
Search grounding
Not supportedSource for Search grounding (opens in a new tab)
Documented tool behavior
This version does not support the image-generation web_search feature.Source for Documented tool behavior (opens in a new tab)
Visible watermark
watermark toggles the visible AI-generated mark.Source for Visible watermark (opens in a new tab)
Batch processing
Not collected

Image sets and completed-image streaming are not proof of an asynchronous Batch API.

Endpoint availability
EOM September 24, 2026; EOS November 24, 2026 at 14:00 UTC+8. Recommended replacement: doubao-seedream-4-0-20260415; automated replacement remains to be confirmed.Source for Endpoint availability (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 15 Oct 2026
CategoryUsageRateUnit
Native image generationSuccessful image output1CN¥0.20output image
Reference input add-onReference image input2CN¥0.00reference image

Rate conditions

  1. Native family tariff per successfully produced image; failed or rejected outputs are not charged. Image sets are metered by actual output count.
  2. Input references are free; this is not free output generation.
Pricing source (opens in a new tab)

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

BytePlus ModelArk · standard

seedream-4-0-250828
Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
Not collected
Billing rules
Not collected
Limits & data boundaries
Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

fal

fal-ai/bytedance/seedream/v3/text-to-image
Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
Not collected
Billing rules
Not collected
Limits & data boundaries
Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 15 Oct 2026

Volcengine Ark images

doubao-seedream-4-0-250828
Not verified

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

How it works
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

Minimum reusable input
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

Lifetime & refresh
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

How to enable it
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

Usage fields
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

Billing rules
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

Limits & data boundaries
Not collected

No image-model prompt-cache contract or cached-input tariff established. Output URL retention is not prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 15 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Inferred from the named native hosted API offering and cross-check of ByteDance and ByteDance-Seed public repositories. No matching downloadable checkpoint was found. This is not an explicit published weight-license declaration.

Classification evidence (opens in a new tab)

Lifecycle dates

Volcengine native Seedream closed to new endpoints24 Sept 2026Source (opens in a new tab)
Volcengine native Seedream shutdown scheduled at 14:00 UTC+824 Nov 2026Source (opens in a new tab)

Sources reviewed