OpenAI releases Images 2.6: ChatGPT learns to create GIF animations

Animations can be generated from a text prompt or based on an uploaded image. For instance, users can ask a drawn character to wave. The playback speed of a finished GIF can be adjusted, and the file can be shared. As Kod.ru reports, citing The Verge, viewing and editing individual frames is also available. The service cannot yet edit animated GIFs uploaded by users themselves: this feature works only for animations created within ChatGPT.

According to OpenAI, Images 2.6 is roughly 50% faster than the previous version. During sequential edits, the model better maintains character appearances, product details, and branding elements. Saved references have also been introduced: an uploaded photo can be saved under a name and then used in other chats without having to be re-uploaded. According to the company, ChatGPT and API users now generate more than 4 billion images a week, up from 3 billion in September.

For developers, two models have been released in the API. GPT-Image-2.6 Flare is designed for speed, while GPT-Image-2.6 Sunburst is built for quality and detail consistency during editing. In addition, area masking for precise edits, SVG export, and generation complexity controls have been added.