What Are the 13 Categories in the GPT-Image2 Gallery? Complete I³ Taxonomy Guide

The GPT-Image2 gallery organizes 544 example cases into 13 distinct I³ (I‑cubed) categories ranging from UI design and infographics to photography and historical art, all defined in docs/gallery.md with navigable anchor links.

The freestylefly/awesome-gpt-image-2 repository maintains a comprehensive visual gallery showcasing GPT-Image2 capabilities through a structured I³ classification system. This taxonomy groups 544 curated examples into thirteen high-level visual domains, making it easy to locate specific use cases and prompt techniques.

Understanding the I³ Classification System

The I³ categories (pronounced "I-cubed") derive their name from "Image" with the superscript ³ indicating a three-tier classification hierarchy. According to the source code in docs/gallery.md, this system creates distinct visual domains that cover both commercial design applications and artistic creation workflows.

The 13 GPT-Image2 Gallery Categories

Each category in the gallery features an emoji marker, case count, and dedicated anchor link for direct navigation. Here are the thirteen categories as defined in the repository:

1. UI与界面 (UI and Interface)

User-interface and screen design covering app interfaces, dashboard layouts, and interactive component mockups. This is the largest category with 73 cases. Anchor: #cat-ui.

2. 图表与信息可视化 (Charts and Information Visualization)

Data visualization graphics including statistical charts, flow diagrams, and analytical infographics. This category contains 53 cases. Anchor: #cat-infographic.

3. 海报与排版 (Posters and Layout Design)

Typography-focused compositions and promotional poster layouts with advanced text arrangement techniques. Anchor: #cat-poster.

4. 商品与电商 (Products and E-commerce)

Commercial product photography and marketplace visuals optimized for online retail presentation. Anchor: #cat-product.

5. 品牌与标志 (Brands and Logos)

Brand identity systems including logo designs, corporate identity packages, and visual branding elements. Anchor: #cat-brand.

6. 建筑与空间 (Architecture and Space)

Architectural renderings and interior spatial designs showing structural visualization capabilities. Anchor: #cat-architecture.

7. 摄影与写实 (Photography and Realism)

Hyper-realistic imagery and photographic art that mimics camera-captured scenes with high fidelity. Anchor: #cat-photo.

8. 插画与艺术 (Illustration and Fine Art)

Artistic illustrations spanning various painting styles and creative visual expressions beyond photorealism. Anchor: #cat-illustration.

9. 人物与角色 (Characters and Portraits)

Character design and portrait work featuring human figures, avatars, and persona-based visual assets. Anchor: #cat-character.

10. 场景与叙事 (Scenes and Narratives)

Storytelling compositions and environmental scenes that convey specific moods, events, or narratives. Anchor: #cat-scene.

11. 历史与古风题材 (Historical and Traditional Style)

Period-specific artwork and traditional cultural aesthetics including classical painting styles and historical recreations. Anchor: #cat-history.

12. 文档与出版物 (Documents and Publications)

Document-style graphics and publication layouts including resumes, reports, and editorial designs. Anchor: #cat-document.

13. 其他应用场景 (Other Applications)

Miscellaneous use-cases that fall outside standard classifications, demonstrating edge-case capabilities. Anchor: #cat-other.

Gallery File Structure and Navigation

The category system is implemented across three main markdown files in the repository. The primary index resides in docs/gallery.md, which defines all thirteen categories and provides navigation anchors. The actual case examples are distributed between docs/gallery-part-1.md (cases 1-165) and docs/gallery-part-2.md (cases 166-544), linked through the category anchors.

Each category header follows a consistent markdown pattern using H3 headings with emoji prefixes and case counts, enabling both human readability and programmatic parsing.

Extracting Category Data Programmatically

You can parse the category metadata directly from the gallery source using Node.js:

import fs from 'fs';
import path from 'path';

const galleryPath = path.resolve(
  __dirname,
  '../docs/gallery.md'
);
const markdown = fs.readFileSync(galleryPath, 'utf-8');

// Regex to capture category headings with case counts
const categoryRegex = /###\s+(.+?)\s+·\s+\d+\s+cases/g;
let match;
const categories = [];

while ((match = categoryRegex.exec(markdown)) !== null) {
  categories.push(match[1].trim());
}

console.log(categories);
// Returns: [
//   'UI与界面',
//   '图表与信息可视化',
//   '海报与排版',
//   '商品与电商',
//   '品牌与标志',
//   '建筑与空间',
//   '摄影与写实',
//   '插画与艺术',
//   '人物与角色',
//   '场景与叙事',
//   '历史与古风题材',
//   '文档与出版物',
//   '其他应用场景'
// ]

This script reads docs/gallery.md and extracts all thirteen category names using a regular expression that matches the H3 heading pattern and associated case statistics.

Summary

• The freestylefly/awesome-gpt-image-2 repository organizes 544 GPT-Image2 examples into 13 I³ categories.
• Categories range from UI design and infographics to photography and historical art.
• Navigation anchors (e.g., #cat-ui, #cat-photo) enable direct linking to specific sections in docs/gallery.md.
• The taxonomy covers commercial applications (e-commerce, branding) alongside artistic domains (illustration, narrative scenes).
• Category metadata can be extracted programmatically using regex patterns against the gallery markdown source.

Frequently Asked Questions

What does I³ stand for in the GPT-Image2 gallery?

I³ (I-cubed) stands for "Image" with a three-tier classification hierarchy. The notation indicates that each category represents a high-level visual domain within the gallery's organizational structure, as implemented in the repository's docs/gallery.md file.

How many example cases are included in the GPT-Image2 gallery?

The gallery contains 544 curated example cases distributed across the thirteen categories. Each category displays its specific count in the heading (ranging from UI with 73 cases to smaller specialized categories), with the full dataset spanning docs/gallery-part-1.md and docs/gallery-part-2.md.

Can I link directly to a specific category in the gallery?

Yes. Each of the 13 categories has a dedicated HTML anchor link defined in docs/gallery.md. For example, the Photography category uses #cat-photo, while Brand identities use #cat-brand. These anchors enable direct browser navigation to specific visual domains without scrolling through all 544 cases.

Is the category list available in languages other than Chinese?

The primary category names in docs/gallery.md use Chinese nomenclature (e.g., "UI与界面", "摄影与写实"), but each entry includes English descriptions in the repository documentation. The bilingual approach serves both Chinese-speaking developers and international users reviewing the GPT-Image2 capabilities.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}