Native animated GIF output
Generate a looping GIF from a prompt or a still image: 2–16 frames, 20–2000 ms per frame, fixed camera. No separate encoder or video conversion.
GPT Image 2.6 Image and GIF Model
GPT Image 2.6 adds native GIF and SVG output, a planning control, and faster generation. Edit a still image with it in the workspace below.
GPT Image 2.6 workspace
0 / 1000
62 credits
The preview image reuses image editing assets and is not GPT Image 2.6 output.
GPT Image 2.6 at a glance
OpenAI's image generation and editing family, offered as Flare for fast iteration and Sunburst for detailed, consistent edits. It keeps the still-image workflow of 2.5 and adds motion, vector output, and a separate planning control.
Read the official model documentationAbout GPT Image 2.6
The model documentation covers text and image-guided GIF output, SVG export, markup-guided edits, and a complexity setting. EzVideo will show only verified modes once the GIF workflow is connected.
What GPT Image 2.6 adds
Native motion and vector output, a separate planning control, and faster generation. These are model capabilities, not features this preview can output yet.
Flare for speed, Sunburst for detail
Flare favors fast iteration and stronger layout for posters and product shots. Sunburst spends more time on detail and consistency.
Around 50% faster generation
OpenAI reports generation roughly 50% faster than the previous generation, with Flare reaching low-quality images in as little as six seconds.
Consistent characters and products
Sunburst at medium or high quality keeps faces, product shapes, logos, and branding closer to the source across a series of edits.
Complexity control
A planning level from low to high decides how much the model plans before drawing, separate from output quality. Use low for focused edits and higher for dense text or crowded scenes.
SVG export and markup edits
Request static SVG artwork for logos and print, or send a transparent markup layer that shows exactly which area should change.
GPT Image 2.6 vs 2.5
Both are OpenAI image models. The table lists what 2.6 adds or changes.
| GPT Image 2.5 | GPT Image 2.6 | |
|---|---|---|
| GIF and SVG output | Not supported | Native GIF (2–16 frames) and SVG |
| Complexity control | Not available | Planning level from low to high |
| Generation speed | Baseline | About 50% faster per OpenAI |
| Pricing | Token-based | Same as 2.5 |
GPT Image 2.6 questions
OpenAI's image generation and editing family with two variants: Flare for fast iteration and Sunburst for detailed, consistent edits.
2.6 adds native GIF and SVG output, a complexity setting, and markup-guided edits in the API, and it generates faster. 2.5 stays a still-image model with the same Flare and Sunburst split and the same token pricing.
Quality controls how much detail is rendered. Complexity controls how much the model plans before drawing and carries no extra charge. Low complexity suits focused edits; medium or high helps with dense text and many objects.
OpenAI token pricing matches GPT Image 2.5. The credit cost shown in the editor depends on the selected model and settings.
Explore GPT Image 2.6 features
Review the model capabilities, then edit a still image with GPT Image 2.6 in the workspace below.