GPT Image 2 Overview
GPT Image 2 is our advanced image synthesis and reference editing engine. It processes complex multi-image reference relationships and subtle character modifications at 1.25 tokens per image render.
Key Capabilities
- Precise Reference Editing: Modify hair styles, clothing, lighting, or expressions while preserving exact facial features from reference photos.
- Colored Pencil & Sketch Transformations: Used internally by Agent Mode for rendering high-precision illustration bypass sheets.
- Complex Spatial Layouts: Renders intricate multi-object scenes with accurate texturing and fine details.
When to Choose GPT Image 2
Use GPT Image 2 when you need exact reference editing, character turnaround sheets, or intricate detail preservation across image-to-image workflows.