OpenAI accelerates and improves image generation with new models
OpenAI has introduced GPT-Image-2.6, significantly expanding image generation capabilities via the API. The new models—Flare and Sunburst—not only accelerate the process, generating images 50% faster, but also improve the quality of iterative editing.
Sunburst is especially useful for sequential editing, as it maintains the detail of characters and objects more stably with each adjustment. Furthermore, users now have the ability to animate static images (sprites, mascots), turning them into looping GIF animations for use in games or applications.
The API now allows for edits to be applied to precisely marked areas of an image, which increases the accuracy and control over the editing process. Additionally, the Codex interface has improved its editing experience, making it more intuitive.
Why it matters
- —Significant increase in generation speed (up to 50%) makes the tool more accessible for commercial use.
- —The introduction of an animation function allows developers to create full media assets (GIFs) from static art.
- —Improved editing accuracy via API, allowing edits to specific, marked areas of an image.
Key facts
- New models released: Image-2.6 Flare and Sunburst.
- Image generation speed increased by 50%.
- Added function to create looping GIFs from static characters.
- API allows edits to precisely specified areas of the image.
The full text is in the original source. Here we provide a brief summary and key facts.