TL;DR
- OpenAI released Images 2.6 on 4 November 2026 in ChatGPT, Codex and the API, saying it generates images 50% faster than before and can turn stills such as mascots or product shots into GIF loops.
- Two API models: GPT-Image-2.6 Flare can generate in as little as six seconds at low quality; GPT-Image-2.6 Sunburst gives higher quality and more consistency across edits at medium or high quality.
- The API also gains image markup, SVG export and a complexity parameter; OpenAI recommends low complexity for most edits and medium or high for scenes with many elements.
- OpenAI says users create 4 billion images a week; the announcement states no prices and says nothing about Thai text in images.
OpenAI released Images 2.6 on 4 November 2026 in ChatGPT, Codex and its API. The update generates images 50% faster than before, adds GIF animation so a still image such as a mascot or product shot can become a loop, and gives developers two API models, GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst, plus image markup, SVG export and a new "complexity" parameter. For marketing teams, it is a faster way to produce ad and social creative, which also means more AI-generated material that someone has to check before it runs.
What OpenAI announced on 4 November 2026
In its Introducing Images 2.6 announcement, OpenAI described the release as faster iteration with more control over editing rounds. The company said users can create polished visuals faster and refine them with greater precision, keeping characters, products and brand details consistent as they edit. It also gave a usage figure: people now create 4 billion images per week across ChatGPT, Codex and the API, which OpenAI put at more than half a billion images per day on average.
On availability, OpenAI said Images 2.6 is available globally in ChatGPT and Codex across all subscription tiers on desktop, mobile and web, and that GPT-Image-2.6 Sunburst and GPT-Image-2.6 Flare are available in the API. The post links to a separate pricing page but does not state prices in the announcement itself, so this article does not quote any.
The features, side by side
The table below lists each capability exactly as OpenAI's 4 November 2026 post describes it.
| Feature | What OpenAI said |
|---|---|
| Speed | Images 2.6 generates images 50% faster than before. |
| GPT-Image-2.6 Flare (API) | Can generate in as little as six seconds with quality set to low; improves layout quality for posters, product imagery and other professional visuals. |
| GPT-Image-2.6 Sunburst (API) | Higher-quality generation and greater consistency across edits with quality set to medium or high. |
| GIFs | Animate a brand mascot, turn a product shot into a loop, make reactions and memes, or build animated storyboards before video production. |
| API controls | Image markup comes to the API, SVG export is added, and a new complexity parameter sets how much planning the model does. |
Two API models: Flare for speed, Sunburst for consistency
The split between the two API models is the most practical part of the release for teams that build creative tools or automate production. OpenAI positions Flare as the fast option: at low quality it can return an image in as little as six seconds, and OpenAI says it improves layout quality for posters, product imagery and other professional visuals. Sunburst is the option for work that needs more detail and precision, with higher-quality output and, in OpenAI's words, even greater image consistency across edits when quality is set to medium or high.
That framing maps onto a familiar creative workflow. Early rounds, where a team is exploring layouts, crops or concepts, reward speed. Final rounds, where the product has to look the same in every variant and the brand details must not drift, reward consistency. OpenAI did not publish benchmark comparisons between the two models in the announcement, so the choice between them is something each team will need to test on its own briefs.
GIFs from still images
The GIF feature is the change most visibly aimed at marketers. OpenAI wrote that marketers and creators can animate a brand mascot for a campaign, create reactions and memes for social media, or use animated storyboards to explore character movement before video production. The post also says a product shot can become an eye-catching loop, and that even a simple sketch can become an animated character for videos, games and more.
The examples OpenAI showed were made by OpenArt, whose CEO and co-founder Coco Mao is quoted in the post describing GIFs as the native language of social media. The prompts OpenArt used are published in full on OpenAI's page and show how the result is built: the model is asked for a sprite sheet of six sequential frames in a three-by-two grid, with a fixed camera, the same character and background in every frame, and a final frame that leads back into the first so the GIF loops.
That detail matters for anyone planning to use the feature in ads. It shows that frame-to-frame consistency is specified in the prompt, not assumed. OpenAI also mentioned Town, a startup building an AI work assistant, which used Images 2.6 in the API to explore how animated characters could add personality to everyday interactions at work.
More control in the API: markup, SVG and complexity
OpenAI listed three additions for developers. Image markup, which was already available in ChatGPT, now comes to the API so teams can pinpoint exactly where to make specific changes. SVG export makes designs ready for websites, apps and print. A new complexity parameter lets developers adjust how much planning the model does to meet a request.
OpenAI separated complexity from quality. In its description, image quality controls visual detail and fidelity, while complexity guides how much goes into the scene. Its recommendation was low complexity for most editing tasks, for faster edits, paired with whatever image quality the workflow needs, and medium or high complexity for scenes with many elements or constraints.
Itay Schiff, Creative Director at Figma Weave, is quoted saying the model will be added to Figma Weave. Davit Avoyan, Director GTM at Picsart, said creators value fast generation and that the model brings more accurate results, better character consistency and faster generations. OpenAI also showed work from early-access customers Picsart, fal, Sable and Higgsfield.
What this changes for ad creative production
The following is analysis of how the announced features fit into creative work, not a claim from OpenAI.
Paid social runs on volume. A single campaign can need the same product in several aspect ratios, multiple hooks and fresh variants once early versions fatigue. Faster generation, a fast model for drafts and a consistency-focused model for finals all reduce the time between a brief and a testable set of variants. The GIF feature adds a cheap motion format between a static image and a full video, which suits placements where a short loop may hold attention better than a still.
SVG export is a quieter change with practical value. Vector output can be edited in standard design tools and scaled cleanly, so a generated icon or illustration can move into a content and design workflow without being redrawn. Image markup in the API means an automated pipeline can request a targeted fix to one region of an image rather than regenerating the whole thing, which is how human designers already work.
None of this removes the need for creative judgement. Faster generation produces more options; it does not decide which of them is accurate, on brand or suitable to run.
The advertiser still owns the ad
This section is general guidance, not something stated in OpenAI's post. When an image goes into an ad, the brand running it is responsible for what it shows, whichever tool made it. AI-generated creative raises specific checks:
- Product accuracy. If a generated or edited image changes a product's shape, colour, packaging or size, the ad may misrepresent what the customer will receive. OpenAI's emphasis on consistency across edits helps, but each final asset still needs a human comparison with the real product.
- Claims and text in the image. Any words, prices or offers that appear inside a generated image need the same review as ad copy.
- Rights in the inputs. Animating a mascot or editing a product shot starts from an image someone supplies. The team should be sure it has the right to use that image and any people or characters in it.
- Platform rules. Ad platforms set their own policies on AI-generated or altered media. Check the current policy of each platform you buy on before launch, because the OpenAI announcement does not cover them.
OpenAI pointed readers to a system card for its safety evaluations and safeguards for Images 2.6. The announcement itself does not describe watermarking, content credentials or commercial usage terms, so teams that need those answers should read the system card and OpenAI's terms directly. The system card, also dated 4 November 2026, says Images 2.6 output carries C2PA metadata and an invisible SynthID watermark across ChatGPT, Codex and the API, and that the GIF feature has its own safeguards, which OpenAI says may not catch every motion-related violation.
What OpenAI did not say
- Prices. The post links to a pricing page but states no figures in the announcement.
- Thai-language text. The post says nothing about rendering Thai script inside images, so Thai teams should test it before relying on it for on-image copy.
- GIF specifications. No maximum length, frame count, resolution or file size is given for GIF output.
- Benchmarks. The 50% speed figure is compared with "before"; the post gives no test conditions and no quality benchmarks between Flare and Sunburst.
- Country detail. OpenAI said "available globally" for ChatGPT and Codex but did not list countries, and did not name Thailand specifically.
What this means for Thai marketers
OpenAI says the ChatGPT and Codex release is global and covers all subscription tiers, so Thai teams on any ChatGPT plan should be able to try GIF loops and faster edits without moving to the API. The most direct uses are motion variants of existing static ads, mascot animations for social ad campaigns and quick storyboards before a video shoot. Two cautions apply locally: test Thai text in images before using it, and keep a human review step for product accuracy and claims, since the brand, not the tool, answers for the ad.
FAQ: OpenAI Images 2.6
What is new in OpenAI Images 2.6?
Images 2.6 generates images 50% faster, adds GIF animation, and introduces two API models, GPT-Image-2.6 Flare and Sunburst, according to OpenAI's 4 November 2026 announcement. Developers also get image markup, SVG export and a complexity parameter in the API.
What is the difference between GPT-Image-2.6 Flare and Sunburst?
Flare is the faster model and Sunburst is the higher-quality, more consistent one. OpenAI said Flare can generate in as little as six seconds at low quality and improves layout for posters and product imagery, while Sunburst gives higher quality and greater consistency across edits at medium or high quality.
Is Images 2.6 available in Thailand?
OpenAI said Images 2.6 is available globally in ChatGPT and Codex across all subscription tiers, but it did not list individual countries. The announcement does not mention Thailand by name or say anything about Thai-language text in images.
How much does GPT-Image-2.6 cost in the API?
The announcement does not state prices. OpenAI links to a separate pricing page, which is the place to check current rates.
Can I use Images 2.6 GIFs in ads?
OpenAI presents GIFs for campaign mascots, social reactions and product loops, but it does not address ad platform rules. The advertiser is responsible for the ad, so check each platform's policy on AI-generated media and review every asset for accuracy before it runs.
If you are planning to bring AI image tools into your creative production, Relevant Audience's content marketing team can help you set up a workflow with the right review steps.
