GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset

2025-07-28

AI Summary

The opacity of closed-source large multimodal models (e.g., GPT-4o) and their proprietary training data severely hinders reproducible research in instruction-guided image editing. Method: This work introduces Edit-1.5M—the first large-scale, publicly available instruction-guided image editing dataset—comprising 1.5 million high-quality “instruction–source image–edited image” triplets. To ensure consistency and alignment fidelity, we innovatively leverage GPT-4o to uniformly reconstruct three major benchmarks (OmniEdit, HQ-Edit, UltraEdit) via image regeneration and semantic instruction rewriting. Contribution/Results: Fine-tuning the open-source FluxKontext model on Edit-1.5M achieves a score of 7.24 on GEdit-EN—significantly narrowing the performance gap with closed-source counterparts. Edit-1.5M establishes a foundational, open, and reproducible resource for instruction-guided image editing research, enabling both data-driven advancement and transparent model evaluation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Recent advancements in large multimodal models like GPT-4o have set a new standard for high-fidelity, instruction-guided image editing. However, the proprietary nature of these models and their training data creates a significant barrier for open-source research. To bridge this gap, we introduce GPT-IMAGE-EDIT-1.5M, a publicly available, large-scale image-editing corpus containing more than 1.5 million high-quality triplets (instruction, source image, edited image). We systematically construct this dataset by leveraging the versatile capabilities of GPT-4o to unify and refine three popular image-editing datasets: OmniEdit, HQ-Edit, and UltraEdit. Specifically, our methodology involves 1) regenerating output images to enhance visual quality and instruction alignment, and 2) selectively rewriting prompts to improve semantic clarity. To validate the efficacy of our dataset, we fine-tune advanced open-source models on GPT-IMAGE-EDIT-1.5M. The empirical results are exciting, e.g., the fine-tuned FluxKontext achieves highly competitive performance across a comprehensive suite of benchmarks, including 7.24 on GEdit-EN, 3.80 on ImgEdit-Full, and 8.78 on Complex-Edit, showing stronger instruction following and higher perceptual quality while maintaining identity. These scores markedly exceed all previously published open-source methods and substantially narrow the gap to leading proprietary models. We hope the full release of GPT-IMAGE-EDIT-1.5M can help to catalyze further open research in instruction-guided image editing.

Problem

• Lack of open-source large-scale image-editing datasets
• Proprietary models limit research accessibility
• Need for high-quality instruction-aligned image triplets

Innovation

• Leverages GPT-4o to unify image datasets
• Enhances images via regeneration and prompt rewriting
• Fine-tunes models for superior editing performance