OpenAI Launches Images 2.6 With GIF Animations and Flare Model

OpenAI has announced the launch of Images 2.6, its next-generation image generation tool, which currently processes four billion images a week. The latest iteration introduces advanced capabilities designed to transform ChatGPT into a comprehensive creative suite, featuring short animations, precise editing tools, and consistent character preservation.

Among the most notable additions is the ability to generate GIFs directly from text prompts or existing images. Users can take a standard photograph, prompt the system to introduce motion, and convert it into a short animation without relying on external editing software.

According to OpenAI, image generation speeds are now up to 50% faster compared to the previous generation. A newly introduced model, designated as Flare, is capable of rendering an image in approximately six seconds under specific conditions, though actual processing times depend on prompt complexity and requested quality.

Another significant enhancement addresses image editing precision. Previously, modifying a minor detail in an AI-generated image often resulted in unintended alterations across the entire composition. The new model ensures that modifications to specific elements, colors, and backgrounds preserve the original characters and design integrity.

Furthermore, users can now save a character or reference image for reuse in future creations, a feature particularly beneficial for producing image series, comic strips, and social media content.

Alongside consumer-facing updates, OpenAI introduced two distinct models for developers: Flare, optimized for speed, and Sunburst, tailored for visual quality and consistency. The company is also adding support for vector graphics export in SVG format via the API.

Images 2.6 is rolling out across ChatGPT, Codex, and the API, with certain capabilities being deployed gradually. Consequently, not all users will immediately access the full suite of new features.