OpenAI has upgraded the image engine in ChatGPT to Images 2.6, introducing animated GIF creation from text prompts and sketches. The model cuts standard rendering times by 50% while improving visual consistency for character faces and logos during targeted edits. The software is available across ChatGPT and Codex, including free accounts and developer APIs.
The update allows users to turn static sketches, photos, or descriptive text into animated GIF clips. Creators can adjust playback speeds and modify specific frames within the finished animation. Standard picture creation runs 50% faster. The engine also maintains identity details across repeated edits, preserving logos and faces when modifying surrounding scenery.
OpenAI split the developer API into 2 distinct builds. GPT Image 2.6 Flare prioritizes speed, rendering draft graphics in 6 seconds. The Sunburst version focuses on complex edits, prompt accuracy, and higher visual quality. Both API tiers support precise inpainting area selection and direct SVG file exports. Weekly output now exceeds 4 billion images, up from 3 billion in previous months.
