由 Google DeepMind 开发并在 Google I/O 2025 大会上推出的 Google Imagen 4 是一款行业领先的文生图模型。它能将提示词转化为具备丰富细节与极高创作自由度的逼真高质量图像。其增强版本 Google Imagen 4 Ultra 在精度、速度和分辨率上实现了进一步突破,使得这两款模型成为商业、营销、设计及创意应用的绝佳选择。

Model Type:
Pricing: Imagen 4 Fast costs 4 credits per image (≈ $0.020). High-tier top-ups (+10% bonus) bring effective pricing down to ≈ $0.018 per image.
Input

The text prompt describing what you want to see

0/5000

A description of what to discourage in the generated images

0/5000

The aspect ratio of the generated image

Random seed for reproducible generation

A configurable parameter. Defaults to true in the Playground.

There is no guarantee that everything can be filtered out; if you are not satisfied with the results, you will need to make your own arrangements.
Output
output typeimage
Generated output
Examples

Explore different use cases and parameter configurations

README

Complete guide to using google/imagen4-fast

Google Imagen 4 API:精准度无与伦比的极速文生图生成

Original image
Imagen 4 Ultra API 的照片级真实渲染
谷歌 Imagen 4 API 的抽象与艺术风格
谷歌 Imagen 4 API 的改进文字排版
Google Imagen 4 Ultra API:电影级与高概念设计
模型Elo 评分(越高越好)预估延迟(秒)相对胜率核心优势与应用场景
Imagen 4 Ultra~1145 (Top performer)~10 (Ultra-fast)Highest overallHighest fidelity, cinematic detail, superb for high-impact visual workflows
Imagen 4~1100~30–35Very highBalanced performance, excellent typography and style versatility
GPT-Imaginative Image (Med/High)~1100–1120~50StrongCreative versatility, good across varied prompt types
DALL·E 3~1030–1050~20–30Moderate-highStrong prompt fidelity, accessible via platforms like ChatGPT
Midjourney v6/v7~1000–1020~40–50ModerateArtistic stylization, widely used for conceptual imagery
Stable Diffusion XL~960–980~50+LowerOpen-source flexibility, customizable workflows
Flux 1.1 Pro / FLUX 1.1~1000+ (varies)~15–20CompetitiveFast with high photo-realism and detailed prompt handling
Seemdream 3 / Recraft V3~960–1000~20–30ModerateBalanced realism with creative flair
4.8/ 5
35,969 people love it
Show some love
  • 01
  • 02
  • 03
  • 04