# Seedream 4.0

> ByteDance's image generation model with integrated text-to-image and image editing capabilities in a unified architecture, supporting up to 4K resolution

## Quick Reference

- Model ID: seedream-4
- Creator: ByteDance
- Status: active
- Family: seedream
- Base URL: https://api.lumenfall.ai/openai/v1

## Specifications

- Max Resolution: 4096x4096
- Input Modalities: text, image
- Output Modalities: image

## Model Identifiers

- Primary Slug: seedream-4


## Tags

image-generation, text-to-image, image-editing

## Available Providers

### fal.ai

- Config Key: fal/seedream-4
- Provider Model ID: fal-ai/bytedance/seedream/v4/text-to-image
- Pricing: $0.030/image
  - Source: https://fal.ai/models/fal-ai/bytedance/seedream/v4/text-to-image

### Replicate

- Config Key: replicate/seedream-4
- Provider Model ID: bytedance/seedream-4
- Pricing: $0.030/image
  - Source: https://replicate.com/bytedance/seedream-4


## Arena Benchmarks

### Studio Ghibli Anime Style

- Elo: 1216
- Record: 15W / 6L / 0T (21 battles)
- Rank: #3 of 13

### Victorian Greenhouse Oasis

- Elo: 1212
- Record: 9W / 7L / 0T (16 battles)
- Rank: #4 of 17

### Isometric Miniature Diorama Scenes

- Elo: 1205
- Record: 8W / 17L / 1T (26 battles)
- Rank: #6 of 19

### Intricate Floral Mandala

- Elo: 1187
- Record: 6W / 7L / 1T (14 battles)
- Rank: #7 of 15

### Over-the-top cartoon caricature

- Elo: 1184
- Record: 8W / 16L / 0T (24 battles)
- Rank: #7 of 13

### Geometric Composition

- Elo: 1180
- Record: 9W / 7L / 2T (18 battles)
- Rank: #7 of 22

### Candid Street Photography

- Elo: 1164
- Record: 9W / 6L / 1T (16 battles)
- Rank: #11 of 22

### Adorable Baby Animals in Sunny Meadow

- Elo: 1164
- Record: 10W / 6L / 1T (17 battles)
- Rank: #10 of 23

### Vintage Cafe Logo

- Elo: 1163
- Record: 12W / 15L / 0T (27 battles)
- Rank: #8 of 19

### Night Sky Transformation

- Elo: 1150
- Record: 10W / 9L / 1T (20 battles)
- Rank: #9 of 15

### Man and Car in California

- Elo: 1142
- Record: 4W / 72L / 1T (77 battles)
- Rank: #13 of 13

### Heroic Super Hero Portrait

- Elo: 1135
- Record: 10W / 16L / 0T (26 battles)
- Rank: #13 of 19

### Modern Clean Menu

- Elo: 1114
- Record: 3W / 19L / 0T (22 battles)
- Rank: #19 of 19

### Bald man challenge

- Elo: 1112
- Record: 4W / 19L / 2T (25 battles)
- Rank: #13 of 14

### Neutral Expression to Genuine Smile

- Elo: 1105
- Record: 8W / 9L / 5T (22 battles)
- Rank: #13 of 13

### Apollo 11: Journey to Tranquility

- Elo: 1102
- Record: 4W / 7L / 0T (11 battles)
- Rank: #15 of 19

### Selective Object Swap

- Elo: 1063
- Record: 1W / 2L / 1T (4 battles)
- Rank: #4 of 12

### Golden Hour Stroll

- Elo: 1048
- Record: 2W / 13L / 0T (15 battles)
- Rank: #12 of 12


## Use Cases & Category Performance

### Anime (Image Editing)

- Rank: #2 of 13
- Elo: 1224
- Record: 15W / 6L / 0T (21 battles)
- Win Rate: 71.4%

### Photorealism (Text-to-Image)

- Rank: #7 of 22
- Elo: 1200
- Record: 9W / 6L / 1T (16 battles)
- Win Rate: 56.3%

### Product, Branding & Commercial (Text-to-Image)

- Rank: #9 of 19
- Elo: 1177
- Record: 12W / 15L / 0T (27 battles)
- Win Rate: 44.4%

### Portrait (Image Editing)

- Rank: #12 of 14
- Elo: 1173
- Record: 11W / 27L / 5T (43 battles)
- Win Rate: 25.6%

### Text Rendering (Text-to-Image)

- Rank: #19 of 21
- Elo: 1190
- Record: 19W / 41L / 0T (60 battles)
- Win Rate: 31.7%

### Photorealism (Image Editing)

- Rank: #15 of 16
- Elo: 1168
- Record: 27W / 121L / 7T (155 battles)
- Win Rate: 17.4%


## Image Gallery

22 images available for this model.
- Curated examples: 4
  - "Cinematic wide shot of a high-end, boutique camera store at twilight. The shop window features a sleek, minimalist ae..."
  - "A hyper-realistic close-up of an artisan watchmaker's desk, scattered with intricate brass gears, tiny springs, and d..."
  - "A hyper-realistic close-up of an artisan watchmaker's workbench, shallow depth of field focusing on a skeletonized me..."
  - "A sunlit, minimalist florist shop with a large chalkboard sign that reads "BLOOMS & BEYOND" in elegant cursive. A var..."
- Competition results: 18
  - Studio Ghibli Anime Style: #3 of 13 (Elo 1216)
  - Victorian Greenhouse Oasis: #4 of 17 (Elo 1212)
  - Isometric Miniature Diorama Scenes: #6 of 19 (Elo 1205)
  - Intricate Floral Mandala: #7 of 15 (Elo 1187)
  - Over-the-top cartoon caricature: #7 of 13 (Elo 1184)
  - Geometric Composition: #7 of 22 (Elo 1180)
  - Candid Street Photography: #11 of 22 (Elo 1164)
  - Adorable Baby Animals in Sunny Meadow: #10 of 23 (Elo 1164)
  - Vintage Cafe Logo: #8 of 19 (Elo 1163)
  - Night Sky Transformation: #9 of 15 (Elo 1150)
  - Man and Car in California: #13 of 13 (Elo 1142)
  - Heroic Super Hero Portrait: #13 of 19 (Elo 1135)
  - Modern Clean Menu: #19 of 19 (Elo 1114)
  - Bald man challenge: #13 of 14 (Elo 1112)
  - Neutral Expression to Genuine Smile: #13 of 13 (Elo 1105)
  - Apollo 11: Journey to Tranquility: #15 of 19 (Elo 1102)
  - Selective Object Swap: #4 of 12 (Elo 1063)
  - Golden Hour Stroll: #12 of 12 (Elo 1048)

## Example Prompt

The following prompt was used to generate an example image in our playground:

A sunlit, minimalist florist shop with a large chalkboard sign that reads "BLOOMS & BEYOND" in elegant cursive. A variety of colorful tulips are in the foreground. In the blurry background, a capybara calmly rests on a terracotta floor tile.

## Code Examples

### Text to Image (Generation)

#### cURL

curl -X POST \
  https://api.lumenfall.ai/openai/v1/images/generations \
  -H "Authorization: Bearer $LUMENFALL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedream-4",
    "prompt": "A serene mountain landscape at sunset",
    "size": "1024x1024"
  }'

# Response:
# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

#### JavaScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.lumenfall.ai/openai/v1'
});

const response = await client.images.generate({
  model: 'seedream-4',
  prompt: 'A serene mountain landscape at sunset',
  size: '1024x1024'
});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
console.log(response.data[0].url);

#### Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.lumenfall.ai/openai/v1"
)

response = client.images.generate(
    model="seedream-4",
    prompt="A serene mountain landscape at sunset",
    size="1024x1024"
)

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
print(response.data[0].url)

### Image Editing

#### cURL

curl -X POST \
  https://api.lumenfall.ai/openai/v1/images/edits \
  -H "Authorization: Bearer $LUMENFALL_API_KEY" \
  -F "model=seedream-4" \
  -F "image=@source.png" \
  -F "prompt=Add a starry night sky to this image" \
  -F "size=1024x1024"

# Response:
# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

#### JavaScript

import OpenAI from 'openai';
import fs from 'fs';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.lumenfall.ai/openai/v1'
});

const response = await client.images.edit({
  model: 'seedream-4',
  image: fs.createReadStream('source.png'),
  prompt: 'Add a starry night sky to this image',
  size: '1024x1024'
});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
console.log(response.data[0].url);

#### Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.lumenfall.ai/openai/v1"
)

response = client.images.edit(
    model="seedream-4",
    image=open("source.png", "rb"),
    prompt="Add a starry night sky to this image",
    size="1024x1024"
)

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
print(response.data[0].url)


## About

## Overview
Seedream 4.0 is a high-resolution image generation and editing model developed by ByteDance. It utilizes a unified architecture that handles both text-to-image synthesis and sophisticated image manipulation within the same framework. This version is specifically designed for high-fidelity outputs, supporting image generation at resolutions up to 4K.

## Strengths
*   **High-Resolution Output:** Demonstrates superior performance in generating 4K imagery, maintaining sharpness and fine textures that often degrade in standard 1024px models.
*   **Unified Task Processing:** Unlike models that require separate adapters (like ControlNet) for editing, Seedream 4.0 integrates generation and editing capabilities into a single pipeline, improving consistency between original and edited pixels.
*   **Multi-Modal Conditioning:** Processes both text prompts and image inputs simultaneously, allowing for precise instruction-following when modifying existing compositions or styles.
*   **Instruction Adherence:** Shows high accuracy in translating complex descriptive prompts into visual elements, particularly in maintaining spatial relationships between objects.

## Limitations
*   **Inference Latency:** Generating at 4K resolution requires significant computational resources, leading to slower generation times compared to lower-resolution models like SDXL or Flux Schnell.
*   **Operational Cost:** At a starting price of $0.03 per generation, it carries a higher cost-per-image than many base latent diffusion models, making it less ideal for high-volume batch processing of simple assets.
*   **Prompt Sensitivity:** While capable of high fidelity, the model’s unified architecture may occasionally require more specific prompting to distinguish between "generate new content" and "preserve existing content" during editing tasks.

## Technical Background
Seedream 4.0 is built on an evolution of ByteDance’s diffusion research, moving away from fragmented architectures toward a unified sequence-to-sequence approach for visual tasks. It treats image editing and generation as a singular objective, which likely involves a high-capacity transformer-based backbone optimized for large-scale spatial parameters. The training regimen emphasizes high-resolution data scaling to ensure the model maintains coherence at 4K dimensions.

## Best For
Seedream 4.0 is ideal for professional design workflows where high-print quality or large-format digital assets are required. It excels in scenarios involving "inpainting" and complex image-to-image transformations where the user needs to maintain the integrity of a base image while adding specific details.

Seedream 4.0 is available through **Lumenfall’s unified API and playground**, allowing developers to integrate its 4K generation and editing capabilities without managing complex infrastructure.

## Frequently Asked Questions

### How much does Seedream 4.0 cost?

Seedream 4.0 starts at $0.03 per image through Lumenfall. Pricing varies by provider. Lumenfall does not add any markup to provider pricing.

### How do I use Seedream 4.0 via API?

You can use Seedream 4.0 through Lumenfall's OpenAI-compatible API. Send requests to the unified endpoint with model ID "seedream-4". Code examples are available in Python, JavaScript, and cURL.

### Which providers offer Seedream 4.0?

Seedream 4.0 is available through fal.ai and Replicate on Lumenfall. Lumenfall automatically routes requests to the best available provider.

### What is the maximum resolution for Seedream 4.0?

Seedream 4.0 supports images up to 4096x4096 resolution.

## Links

- Model Page: https://lumenfall.ai/models/bytedance/seedream-4
- About: https://lumenfall.ai/models/bytedance/seedream-4/about
- Providers, Pricing & Performance: https://lumenfall.ai/models/bytedance/seedream-4/providers
- API Reference: https://lumenfall.ai/models/bytedance/seedream-4/api
- Benchmarks: https://lumenfall.ai/models/bytedance/seedream-4/benchmarks
- Use Cases: https://lumenfall.ai/models/bytedance/seedream-4/use-cases
- Gallery: https://lumenfall.ai/models/bytedance/seedream-4/gallery
- Playground: https://lumenfall.ai/playground?model=seedream-4
- API Documentation: https://docs.lumenfall.ai