OpenAI’s new Images 2.6 release is easy to summarize as a speed upgrade. For artists, the more consequential change is control: the system is designed to preserve characters, products and visual details through repeated edits, while adding animated GIF output and new controls for quality and scene complexity.
That moves generative imaging closer to a production workflow and farther from the one-prompt novelty that defined its public image. The shift matters because professional visual work is rarely about producing one impressive frame. It is about revision, continuity, delivery specifications and the ability to explain what changed between versions.
From image generation to iterative production
OpenAI says Images 2.6 is available globally in ChatGPT and Codex, with two API variants called Flare and Sunburst. The company reports that the model is 50 percent faster than its predecessor. Flare prioritizes speed and can generate low-quality outputs in as little as six seconds, while Sunburst is positioned for greater detail and consistency at medium and high quality settings.
A separate complexity control lets developers adjust how much planning the model applies to a request. That distinction is useful. Image quality governs visual fidelity; complexity governs how much work the system does to satisfy a scene with multiple elements or constraints. For artists, designers and studios, the practical question is no longer only whether a model can make an image, but whether it can maintain a visual system through a sequence of decisions.
The release also introduces GIF generation. OpenAI presents this as a way to animate mascots, make social-media reactions and test movement in storyboards. In a working studio, those uses may extend to proposals, exhibition previews, motion studies and client-facing concept development. The risk is that speed can make provisional material look finished before its authorship, rights and purpose have been examined.
Consistency changes the economics of revision
Character and object consistency have been persistent weaknesses in generative imaging. When those elements drift between versions, artists spend time rebuilding prompts, correcting details or moving the work into conventional editing software. Better continuity can reduce that friction and make hybrid workflows more practical.
It can also change expectations. A client who sees a six-second draft may assume that finished work should arrive almost as quickly. Artists should separate generation time from creative labor: research, art direction, selection, retouching, typography, permissions, accessibility, file preparation and accountability remain human responsibilities even when the first image appears instantly.
A useful production record can be simple. Note the model and date, preserve prompts that materially shaped the result, identify reference files and their permissions, save major iterations, and document the human edits that transformed an output into a finished work. For commissioned projects, contracts should say whether generative tools are permitted, who approves their use, who bears responsibility for supplied references and what disclosures accompany delivery.
What the announcement does not settle
OpenAI’s release describes capabilities, pricing routes and availability, and points readers to a system card for safety evaluations. It does not by itself resolve the legal status of every output, the rights implications of every uploaded reference, or the disclosure rules that may apply in a client’s market. Those questions depend on the material used, the jurisdiction, the contract and the way the finished work is distributed.
The scale is significant: OpenAI says people now create four billion images each week across ChatGPT, Codex and the API. At that volume, consistency and speed are not merely convenience features. They can influence visual culture, production schedules and the price assigned to routine imagery.
For artists, the strongest response is neither automatic adoption nor automatic rejection. Test the tool on a bounded task, compare the time saved with the control surrendered, and preserve enough process evidence to explain the result. The most important creative question remains human: what is this image for, and who is responsible for it?
ATIM approaches artificial intelligence with curiosity, independence, and balance. We explore its extraordinary potential to empower artists while examining the questions it raises around authorship, copyright, consent, creative identity, transparency, and human expression.
We are neither advocates for AI nor opponents of it. Our focus is how the technology is used—and what that means for artists.
Above all, our commitment is to artists. We look at every development through the lens of what can support, protect, and empower the creative community—identifying opportunities, examining potential risks, and providing artists with the information they need to make informed decisions about their work, their rights, and their future.