GPT Image 1.5

🎯 Precise Instruction-Based Image Editing Users can describe desired edits in natural language, and GPT Image 1.5 accurately identifies what to change and what to keep intact, completing edits without disrupting the overall composition.

🧬 Identity & Context Preservation For edits involving people, GPT Image 1.5 actively maintains consistent facial features, expressions, lighting, and overall atmosphere, ensuring the subject still looks like the same person after multiple edits.

📝 In-Image Text Generation & Editing Supports generating clear, readable text in images, including poster copy, product labels, and signage, with proper handling of curved surfaces, perspective, and multilingual characters.

🔁 Multi-Image & Multi-Frame Consistency Ideal for generating image series. Characters, props, and environments are "remembered" to maintain consistent appearance and style across different frames, cutting down on repeated description effort.

🧩 Element Addition & Removal Naturally insert new objects or seamlessly remove unwanted elements from the frame, with automatic background filling to make results look like original photos.

📐 High-Quality Output Supports export up to 4K resolution, with multiple format options (PNG / JPEG / WebP), suitable for display, printing, or commercial delivery.