OpenAI has launched the new Images 2.6 model, enabling ChatGPT and Codex users to create animated GIFs directly from text prompts and uploaded photos across web, desktop, and mobile applications. The update allows users to generate GIFs and subsequently fine-tune the animation speed or edit individual frames. During demonstrations, OpenAI showcased how a prompt for a dancing cat creates multiple variations strung together in a flipbook style, highlighting the model's improved image consistency that keeps characters recognizable across frames compared to previous iterations.
Along with GIF creation, the Images 2.6 model aims to accelerate overall image generation speeds for the platform. For developers, OpenAI is releasing new API model variants including GPT-Image-2.6 Flare for efficient generation and Sunburst for higher-quality output. OpenAI reports that users currently generate well over four billion images every week through ChatGPT and the API, marking a substantial increase of one billion weekly images since September.
This feature rollout significantly expands the multimedia capabilities of conversational AI tools for both everyday users and developers.