OpenAI ships GPT Image 2.6, two models built for speed or detail
Sunburst and Flare split image generation into a quality model and a fast one, and both can now output SVG and animated GIFs straight from the API.
OpenAI has shipped GPT Image 2.6 as two separate models: gpt-image-2.6-sunburst, built for detailed compositions and editing, and gpt-image-2.6-flare, built for fast, interactive generation and editing. Both ship in ChatGPT and through the Images API.
The bigger change is what the API itself can now do. A new complexity parameter (low, medium or high, defaulting to high) controls how much planning the model does before it renders, at no extra cost. Both models can also produce SVG output and animated GIFs directly through the API, rather than you generating a raster image and converting it yourself. Pricing is identical across both models: $5 per million text input tokens ($1.25 cached), $8 per million image input tokens ($2 cached), and $30 per million image output tokens.
{
"model": "gpt-image-2.6-flare",
"complexity": "low"
}Why it matters: if you’re calling the Images API, you now pick a model for the speed-versus-quality tradeoff you want, and SVG or GIF output stops being a conversion step you build yourself.
The caveat: OpenAI’s docs don’t mark gpt-image-2.5 as deprecated, so this is an addition rather than a forced migration, and no shutdown date has been set for the older models.