GPT Image 1.5 for image editing, leverages the base model's "logic-first" approach to offer superior performance in localized editing tasks. Using the OpenAI Image API, it takes an image and a text-based instruction to perform precise in-painting (editing within a masked area) and out-painting (extending the image canvas). It excels at workflows requiring high consistency, such as modifying product details, retouching portraits, or altering specific background elements, all while preserving the original image's lighting, composition, and overall integrity.