OpenAI Images 2.6: faster images, GIFs and API control

What changed?

OpenAI announced Images 2.6, available in ChatGPT, Codex and the API. Generation is about 50% faster than before, and the company says people now create about 4 billion images a week across those three surfaces.

The headline feature is motion: a single prompt can produce a looping GIF for brand mascots, reactions or storyboards. Edits also keep characters, products and brand details more consistent across rounds.

What you can build

• The API exposes GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. At low quality, Flare can return an image in as little as six seconds and improves poster and product layouts.
• Sunburst at medium or high quality aims for richer detail and stronger consistency across edit rounds.
• Image markup, already in ChatGPT, now lands in the API so you can mark exactly where the next change should go. A new complexity parameter steers how much planning the model does, and SVG export carries designs into web, app or print workflows.

Who is using it

OpenArt is shipping GIFs from a single prompt, Town is exploring animated characters for a workplace assistant UI, and Figma Weave is wiring GPT-Image-2.6 into creative production workflows. Images 2.6 is live globally in ChatGPT and Codex on every subscription tier across desktop, mobile and web. Flare and Sunburst ship in the API; see API pricing for rates.

Why it matters

If you iterate on product shots, poster layouts or animated mascots, Images 2.6 packs speed and edit control into one stack. Use Flare for fast low-quality loops and Sunburst when you need medium or high fidelity across edits; complexity and markup should cut the number of rounds. Read the system card before you put it in production.