OpenAI just shipped Images 2.6 across ChatGPT, Codex and the API. For developers, the core release is GPT-Image-2.6, shipped as two API models: Flare and Sunburst. OpenAI team states that Images 2.6 generates images 50% faster than before. The update also adds GIF output, SVG export, markup-guided edits and a new complexity parameter.
TL;DR
- Size: Closed weights, token-billed image API.
- Runs on: OpenAI-hosted only (ChatGPT, Codex, API).
- Performance: 50% faster image generation than the previous release, per OpenAI.
- Best: Flare can return an image in as little as 6 seconds at low quality.
- Worst: No benchmark scores and no per-image cost table published for 2.6.
What is GPT-Image-2.6?
GPT-Image-2.6 is OpenAI’s image generation and editing model family for the API. It accepts text and image inputs and returns images. GPT-Image-2.6 Flare is built for fast iteration and interactive workflows. GPT-Image-2.6 Sunburst is built for detailed compositions and editing.
OpenAI states that Flare can generate images in as little as 6 seconds when quality is set to low. Sunburst targets higher quality and stronger consistency across edits at medium or high quality.
What changed in ChatGPT Images 2.6?
ChatGPT users can now turn a prompt or a photo into an animated GIF. They can adjust animation speed, edit individual frames and share the result. Users can also save and tag reference images of people, pets and characters. Those references can be reused later without uploading them again.
Usage keeps climbing. OpenAI says people now generate more than 4B images per week across ChatGPT and the API. That is up from more than 3B per week 8 weeks earlier, at the Images 2.5 launch. On average, that is over 500M images per day.
How does the new complexity parameter work?
The complexity field sets how much the model plans before it renders. The image generation guide lists 3 levels: low, medium and high. If you omit it, both models use high.
Low suits focused edits, such as changing a background or a sign. High helps with dense text, complex layouts and relationships between objects. OpenAI warns that higher complexity can add latency and introduce unwanted changes. Adjusting complexity has no separate fee. It is separate from quality, which controls rendering detail and fidelity.