Seedream 5.0 Pro API

ByteDance / BytePlus-图像生成已开放

EvoLink 1K 价格 $0.034/张-25%$0.045 official price

文生图图生图多参考编辑商品创意角色一致性商用图像输出

渠道状态

已开放异步任务跟踪

模型亮点

高保真图像生成与编辑

适用场景

广告、社媒内容、商品创意、视觉概念

输入

文本 Prompt，可选图像参考

输出

图像

Playground Pricing About API

接入 API 之前，先试用 Seedream 5.0 Pro。

在正式接入前，测试 Seedream 5.0 Pro 的输出质量、估算 credits 消耗，并生成一张样图。

Create image task

Fixed to doubao-seedream-5.0-pro

Text-to-image, image-to-image and editing. One image per request.

Prompt

Input images

Upload source or reference images. Each input image is billed.

0/10 files

Size

Quality

Output format

Background

Transparent needs exactly one input image that already has an alpha channel.

Estimate $0.0342.295 cr

1K quality + auto size + 0 billable input images

Ready

History

生成结果保留 24 小时，需要的请及时下载。

Generated Seedream 5.0 Pro tasks will appear here.

充值前先算清测试成本。

从单张样图的成本起步，选一个测试预算。展开查看详细费率。

First Test Cost

Text-to-Image - 1K - 1:1

1K prompt-only

Credits2.295

One 1K prompt-only image

Approx. Cost$0.034

About 294 tests with $10 in credits.

Iterate prompts in 1K to save credits, then switch to 2K for the final quality check. Each input image is billed when image_urls is present.

EvoLink 价格$0.034

official price$0.045

节省-25%

Testing Budget Guide

Pick an amount based on how many tests you expect.

Add Credits

$10

About 294 tests

Good for a first validation.

$50

About 1471 tests

Good for prompt iteration.

$100

About 2941 tests

Good before production integration.

Common Cost Examples

Draft iteration1K prompt-only

$0.034/image-25%

2.295 cr/image$0.045 official price

Reference edit1K + one input image

$0.036/image-25%

2.448 cr/image$0.048 official price

Final quality check2K prompt-only

$0.068/image-25%

4.59 cr/image$0.090 official price

Estimated cost for one generated image. Final cost may vary by output quality and input images.

Model Pricing

Item	Rule	Rate	Billing
1K / 1.5K output	Single generated image at the 1K or 1.5K quality tier. Both tiers cost the same, so 1.5K is the better pick.	$0.034/image-25% 2.295 cr/image$0.045 official price	Output image
2K output	Single generated image using the 2K quality tier.	$0.068/image-25% 4.59 cr/image$0.090 official price	Output image
Layer Split · 1K / 1.5K	Each image produced by Layer Split at the 1K or 1.5K tier. The official layer-decomposition rate is half the standard generation rate.	$0.018/image-20% 1.224 cr/image$0.023 official price	Output image
Layer Split · 2K	Each image produced by Layer Split at the 2K tier. A run yields one base image plus up to 16 layers, each priced by its own pixel count.	$0.036/image-20% 2.448 cr/image$0.045 official price	Output image

1K / 1.5K output

Single generated image at the 1K or 1.5K quality tier. Both tiers cost the same, so 1.5K is the better pick. - Output image

$0.034/image-25%

2.295 cr/image$0.045 official price

2K output

Single generated image using the 2K quality tier. - Output image

$0.068/image-25%

4.59 cr/image$0.090 official price

Layer Split · 1K / 1.5K

Each image produced by Layer Split at the 1K or 1.5K tier. The official layer-decomposition rate is half the standard generation rate. - Output image

$0.018/image-20%

1.224 cr/image$0.023 official price

Layer Split · 2K

Each image produced by Layer Split at the 2K tier. A run yields one base image plus up to 16 layers, each priced by its own pixel count. - Output image

$0.036/image-20%

2.448 cr/image$0.045 official price

Optional Add-ons & Billing NotesInput image pricing and live console pricing.

Item	Rule	Rate	Billing
Input images	Applied to each source or reference image in image_urls.	$0.0023/image-25% 0.153 cr/image$0.0030 official price	Input images

Seedream 5.0 Pro 能创造什么。

用文本 Prompt 或可选参考图生成高保真图像，在商品创意、海报排版、角色设定图、视觉概念和图像编辑工作流上表现出色。

更多 Prompt 示例

输入

文本 / 图像

输出

图像

尺寸

自动 / 比例 / 自定义像素

质量

1K / 1.5K / 2K

效果预览

复杂街景与可读文字渲染

街景里的招牌与文字直接生成即可读，无需后期贴字。

样张示例，实际结果会随 Prompt 和参数设置变化

写实人物与多主体群像

同一画面多个人物，表情自然、光线统一。

样张示例，实际结果会随 Prompt 和参数设置变化

多参考商品合成

把多张商品与风格参考合成到一张统一的编排版式里，每个元素都保持品牌一致。

样张示例，实际结果会随 Prompt 和参数设置变化

你可以做什么

生成营销视觉

用一个 Prompt 生成广告、上新海报、电商主图和社媒创意。

用参考图编辑

用 image_urls 引导风格、主体一致性、构图或转换。

控制尺寸与质量

尺寸用自动、比例预设或自定义像素；质量选 1K 或 2K 输出清晰度。

跟踪 credits 消耗

这条按图计费的路由会返回 credit 与图像用量字段。

Seedream 5.0 Pro 与其他图像模型对比

特性	Seedream 5.0 Pro	Nano Banana	GPT Image
起步价	约 $0.034 / 张	更低成本选项	按 token 或档位计费
输出质量	1K / 1.5K / 2K	按模型支持 1K / 2K / 4K	质量档位
参考图输入	最多 10 张	取决于模型	支持
输入图计费	每张输入图计费	取决于路由	按 token 计费
适合场景	广告、商品创意、图像编辑	高质量图像生成	OpenAI 图像工作流

起步价

Seedream 5.0 Pro约 $0.034 / 张

Nano Banana更低成本选项

GPT Image按 token 或档位计费

输出质量

Seedream 5.0 Pro1K / 1.5K / 2K

Nano Banana按模型支持 1K / 2K / 4K

GPT Image质量档位

参考图输入

Seedream 5.0 Pro最多 10 张

Nano Banana取决于模型

GPT Image支持

输入图计费

Seedream 5.0 Pro每张输入图计费

Nano Banana取决于路由

GPT Image按 token 计费

适合场景

Seedream 5.0 Pro广告、商品创意、图像编辑

Nano Banana高质量图像生成

GPT ImageOpenAI 图像工作流

关键细节

尺寸决定版式

选定 1K/1.5K/2K 档位用自动，选 1:1、16:9 等比例预设，或传自定义 WIDTHxHEIGHT 像素。

质量即清晰度

用质量选择输出档位。1K 最适合测试，2K 更适合终稿核对。

每张输入图计费

纯 Prompt 请求只付输出图费率。图像工作流会对 image_urls 中每张源图或参考图计费。

为什么通过 EvoLink 使用 Seedream 5.0 Pro

生成前成本透明

充值前即可看到预估 credits、输入图附加费和样图成本。

统一 API

一把 API key 通用于 Seedream、Nano Banana、GPT Image 以及视频、音频和 LLM 模型。

统一余额

在一个控制台里跟踪 credits、消费和模型用量。

任务状态跟踪

查看图像任务处于已提交、处理中、已完成还是失败。

生产级图像工作流

在一个页面上测试商品广告、社媒内容、角色参考和编辑工作流。

Seedream 模型家族

Seedream 5.0 Lite

成本更低的 Seedream 5.0 路由，适合大范围图像生成和联网搜索增强的创意探索。

查看

Seedream 4.5

上一代 Seedream 图像路由，适合需要 2K/4K 生成、多图编辑和成熟生产行为的团队。

查看

Seedream 4.0

更早的 Seedream 图像模型，比例支持广、文生图工作流稳定。

查看

EvoLink 上的其他图像模型

Nano Banana 2

Google 图像模型，高质量生成、创意工作流灵活。

查看

Nano Banana Pro

高阶 Nano Banana 路由，图像更高保真、输出达生产级。

查看

GPT Image 2

OpenAI 图像路由，支持 Prompt、参考图和编辑工作流。

查看

Qwen Image Edit

图像编辑路由，适合直接转换和局部创意改动。

查看

常见问题

Seedream 5.0 Pro 和 Seedream 5.0 Pro Layerize 有什么区别？

它们是两个独立模型。doubao-seedream-5.0-pro 生成或编辑一张平面图，支持比例预设和自定义像素尺寸。doubao-seedream-5.0-pro-layerize 只收一张 PNG 或 JPEG，尺寸只接受档位，产出一张底图加最多 16 个透明 PNG 图层，并带叠放次序和包围盒。要出图用 Pro，已有构图还要再改用 Layerize。

Seedream 5.0 Pro 多少钱？

按输出质量档位计费。1K 纯 Prompt 图用 1K 输出费率，2K 用更高的输出费率。传入 image_urls 时，输入图按张计费。

接入 API 前能先预览吗？

可以。用 Playground 测试输出质量、估算 credits，并在接入产品前查看请求 JSON。

尺寸和质量有什么区别？

size 控制自动尺寸、比例或自定义像素。quality 控制清晰度档位，当前为 1K 或 2K。

所有输入图都计费吗？

是。image_urls 中每张输入图都会产生输入图费用。

用量包含 token 数吗？

不。面向用户的用量按 credit 与图像计：已用 credits、输入图数量、可计费输入图数量、生成图数量和输出档位。

能用 n 一次生成多张吗？

暂不支持。Seedream 5.0 Pro 当前固定 n=1。

能用于图像编辑吗？

可以。当你想用源图或参考图引导编辑或图生图工作流时，随 Prompt 一起传 image_urls。

生成失败怎么办？

在响应或控制台查看任务状态和错误详情。失败任务不应视为计费成功的生成。

充值后去哪里？

回到 Playground 继续测试，或创建 API key 并用 model doubao-seedream-5.0-pro 调用 /v1/images/generations。

图层拆分深入阅读

Layerize 上线说明

图层输出的完整约定、与 BytePlus 官方价的对比，以及它改变了哪些工作流。

API Reference

Select endpoint

Model ID

Endpoints

Authentication

All APIs require Bearer Token authentication.

Header

Authorization: 
Bearer YOUR_API_KEY

Get API Key

POST

/v1/images/generations

Generate Image

Create an image generation or editing task using text prompts and optional reference images. Supports auto size, 1K/2K output tiers, ratio presets, custom pixel size outputs, and one generated image per request.

Asynchronous processing mode, use the returned task ID to .

Generated image links are valid for 24 hours, please save them promptly.

Request Parameters

modelstringRequiredDefault: doubao-seedream-5.0-pro

Image generation model name.

Value	Description
doubao-seedream-5.0-pro	Seedream 5.0 Pro model
doubao-seedream-5.0-pro-layerize	Layer Split: returns one base image plus up to 16 transparent PNG layers. Half the price. See the notes below for its parameter differences.

Notes

Layer Split (doubao-seedream-5.0-pro-layerize) narrows the contract: image_urls is required and must contain exactly one png or jpeg image; prompt becomes optional (omit it to auto-detect every major element); size/quality accept only auto, 1K, 1.5K or 2K — aspect ratios and WIDTHxHEIGHT are rejected; background, n and model_params are not supported.
Each Layer Split image is billed by its own pixel count, so a run costs far less than the amount reserved upfront.

Exampledoubao-seedream-5.0-pro

promptstringRequired

Prompt describing the image to be generated.

Notes

Up to 4000 tokens (roughly 2,600 Chinese characters or 5,000 English words); longer prompts are rejected with 400 and not billed
Besides Chinese and English, Seedream 5.0 Pro also supports Russian, Arabic, Filipino, Thai, Turkish, Korean, Malay, Spanish, Portuguese, Indonesian, French, German, Vietnamese and Japanese
Recommended: under 300 Chinese characters or 600 English words — longer prompts stay within the limit but dilute the instruction

ExampleA serene lake reflecting the beautiful sunset

sizestringOptionalDefault: auto

Size of generated images. Supports aspect ratios or custom pixel dimensions.

Value	Description
auto	No explicit ratio — the model decides the composition from the prompt, and the output resolution follows the tier set by quality
1:1 · 4:3 · 3:4 · 3:2 · 2:3	Aspect ratio presets — see the quality parameter for the exact pixels each tier resolves to
16:9 · 9:16 · 4:5 · 5:4 · 21:9 · 9:21	Aspect ratio presets — see the quality parameter for the exact pixels each tier resolves to
WIDTHxHEIGHT	Custom pixel dimensions, e.g. 2048x2048

Notes

Use size=auto to let the model pick the composition from the prompt; the resolution then follows the quality tier
Ratio presets resolve to exact pixels according to quality — see that parameter for the full table
Custom WIDTHxHEIGHT: total pixels within [921600, 4624220], aspect ratio within [1/16, 16], and both width and height must be greater than 14px

Exampleauto

qualitystringOptionalDefault: 1K

Output clarity tier for auto or ratio size. With exact WIDTHxHEIGHT, quality is ignored and the actual output pixels decide the billing tier: up to 2,610,000 px bills at the lower tier, above that at the higher tier.

Value	Description
1K	1:1 1024×1024 · 4:3 1152×864 · 3:4 864×1152 · 3:2 1248×832 · 2:3 832×1248 · 16:9 1424×800 · 9:16 800×1424 · 4:5 896×1120 · 5:4 1120×896 · 21:9 1568×672 · 9:21 672×1568
1.5K	1:1 1536×1536 · 4:3 1792×1344 · 3:4 1344×1792 · 3:2 1872×1248 · 2:3 1248×1872 · 16:9 2048×1152 · 9:16 1152×2048 · 4:5 1344×1680 · 5:4 1680×1344 · 21:9 2352×1008 · 9:21 1008×2352
2K	1:1 2048×2048 · 4:3 2368×1776 · 3:4 1776×2368 · 3:2 2496×1664 · 2:3 1664×2496 · 16:9 2816×1584 · 9:16 1584×2816 · 4:5 1792×2240 · 5:4 2240×1792 · 21:9 3136×1344 · 9:21 1344×3136

Notes

1K and 1.5K cost the same and 1.5K looks better, so prefer 1.5K
With the pixel format of size, the tier follows the actual output pixel count: up to 2,610,000 is billed at the lower tier, above that at the higher tier

Example1K

prompt_prioritystringOptionalDefault: standard

Prompt optimization strategy.

Value	Description
standard	Standard mode, higher quality, slower
fast	Fast mode, quicker but slightly lower quality

Examplestandard

watermarkbooleanOptionalDefault: false

Whether to add a watermark to the generated image.

Value	Description
false	Do not add a watermark (default)
true	Add a watermark

Examplefalse

backgroundstringOptionalDefault: opaque

Alpha channel control. transparent keeps the alpha channel your input image already has — it does not cut out an opaque background.

Value	Description
opaque	Regular solid background (default)
transparent	Keep the input image's alpha channel

Notes

transparent requires exactly one input image that already has an alpha channel
transparent forces PNG output; passing output_format=jpeg is rejected
An input image without an alpha channel (for example jpeg) is rejected upstream

Exampleopaque

image_urlsarrayOptional

Reference image URL list for image-to-image and image editing.

Notes

Max 10 images per request
Max file size: 30MB per image
Supported formats: .jpeg, .jpg, .png, .webp, .bmp, .tiff, .gif, .heic, .heif
Aspect ratio (width/height): [1/16, 16]
Total pixels: [196, 6000×6000]
Each input image is billed separately

Example["https://example.com/image1.png"]

output_formatstringOptionalDefault: jpeg

Output image format.

Value	Description
jpeg	Compressed format, smaller file size (default)
png	Lossless format

Examplejpeg

callback_urlstringOptional

HTTPS callback address after task completion.

Notes

Triggered on completion, failure, or cancellation
Sent after billing confirmation
HTTPS only, no internal IPs
Max length: 2048 chars
Timeout: 10s, Max 3 retries

Examplehttps://your-domain.com/webhooks/image-task-completed

Examples

Request Example

{
  "model": "doubao-seedream-5.0-pro",
  "prompt": "Create a polished product poster for a premium electric bicycle",
  "size": "auto",
  "quality": "1K",
  "output_format": "jpeg",
  "image_urls": [
    "https://example.com/reference-product.png"
  ],
  "callback_url": "https://your-domain.com/webhooks/callback"
}

Response Example

{
  "created": 1772000000,
  "id": "task-unified-1772000000-seedream5pro",
  "model": "doubao-seedream-5.0-pro",
  "object": "image.generation.task",
  "progress": 0,
  "status": "processing",
  "task_info": {
    "can_cancel": true,
    "estimated_time": 30
  },
  "type": "image",
  "usage": {
    "billing_rule": "per_image",
    "credits_reserved": 2.448,
    "user_group": "default"
  }
}

Completed Usage Example

{
  "id": "task-unified-1772000000-seedream5pro",
  "object": "image.generation.task",
  "created": 1772000000,
  "model": "doubao-seedream-5.0-pro",
  "status": "completed",
  "type": "image",
  "progress": 100,
  "results": ["https://cdn.evolink.ai/images/generated-image.jpg"],
  "result_data": [
    {
      "url": "https://cdn.evolink.ai/images/generated-image.jpg",
      "size": "1024x1024",
      "output_format": "jpeg"
    }
  ],
  "usage": {
    "credits_used": 2.601,
    "input_images": 2,
    "billable_input_images": 2,
    "generated_images": 1,
    "output_tier": "1K",
    "output_pixels": 1048576
  }
}

Seedream 5.0 Pro API

接入 API 之前，先试用 Seedream 5.0 Pro。

充值前先算清测试成本。

Common Cost Examples

Model Pricing

Seedream 5.0 Pro 能创造什么。

效果预览

复杂街景与可读文字渲染

写实人物与多主体群像

多参考商品合成

你可以做什么

生成营销视觉

用参考图编辑

控制尺寸与质量

跟踪 credits 消耗

Seedream 5.0 Pro 与其他图像模型对比

关键细节

尺寸决定版式

质量即清晰度

每张输入图计费

为什么通过 EvoLink 使用 Seedream 5.0 Pro

生成前成本透明

统一 API

统一余额

任务状态跟踪

生产级图像工作流

Seedream 模型家族

Seedream 5.0 Lite

Seedream 4.5

Seedream 4.0

EvoLink 上的其他图像模型

Nano Banana 2

Nano Banana Pro

GPT Image 2

Qwen Image Edit

常见问题

Seedream 5.0 Pro 和 Seedream 5.0 Pro Layerize 有什么区别？

Seedream 5.0 Pro 多少钱？

接入 API 前能先预览吗？

尺寸和质量有什么区别？

所有输入图都计费吗？

用量包含 token 数吗？

能用 n 一次生成多张吗？

能用于图像编辑吗？

生成失败怎么办？

充值后去哪里？

图层拆分深入阅读

Layerize API 教程

Layerize 上线说明

API Reference

Authentication

Generate Image

Request Parameters

Notes

Notes

Notes

Notes

Notes

Notes

Notes

Examples

Request Example

Response Example

Completed Usage Example