OpenAI Images 2.6 is live in ChatGPT, Codex and the API. In a 4 November 2026 post, OpenAI said the release is available globally across every subscription tier on desktop, mobile and the web, with two API models: GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst.
The launch is less about a single new demo trick and more about control: animated GIF output, SVG export, markup-guided edits in the API, and a new “complexity” setting that sits beside the existing quality control. OpenAI also says people now create about 4 billion images a week across ChatGPT, Codex and the API. That figure, like the speed claims below, comes from OpenAI itself.
In brief
- OpenAI Images 2.6 is live in ChatGPT, Codex and the API for all tiers.
- API users choose between Flare (fast) and Sunburst (higher detail); GIF, SVG and markup edits are new.
- Speed, quality and usage figures are OpenAI’s own; Flare pricing was not checked on the Sunburst docs page.
What OpenAI Images 2.6 Flare and Sunburst offer
In the API, Flare and Sunburst are separate model IDs. OpenAI says Images 2.6 generates images 50% faster than before, and that Flare can return an image in as little as six seconds when quality is set to low. Sunburst is pitched for higher-detail work and stronger consistency across edits when quality is set to medium or high. Flare, OpenAI adds, also improves layout quality for posters and product imagery.
The GPT-Image-2.6 Sunburst model page lists token pricing for that variant: text input at $5 per 1M tokens (cached $1.25), image input at $8 per 1M (cached $2), and image output at $30 per 1M. The same page confirms text and image inputs, image outputs, and support for the image generation and image edit endpoints. Flare’s price is not on that page; OpenAI points developers to its pricing pages for the full picture.
Alongside the two models, Images 2.6 adds tools that matter for production pipelines. Image markup, already in ChatGPT, now reaches the API so teams can mark exactly where a change should land. SVG export is aimed at websites, apps and print. A new complexity parameter steers how much planning the model does: OpenAI recommends low complexity for quick edits, and medium or high when a scene has many constraints.
Motion, partners and availability
GIF output is the most visible consumer change. OpenAI frames it for brand mascots, social reactions and animated storyboards before video work. The post quotes OpenArt’s Coco Mao and Figma Weave’s Itay Schiff praising the model’s fit for social GIFs and reusable creative workflows. Those are customer endorsements, not independent tests.
Availability is broad by OpenAI’s standards: ChatGPT and Codex on all tiers, plus both API models. OpenAI also links a system card for safety evaluations of Images 2.6; TSN has not summarised that card here.
Why image models keep splitting into fast and careful lanes
The Flare and Sunburst split follows a pattern already visible across the image market. Creators want a cheap, quick pass for drafts and a slower, more faithful pass for brand work. Microsoft’s MAI Image 2.7 preview pushed on editing leaderboards, while Canva’s in-house Craft model put generation inside a design product rather than a general chatbot. OpenAI is doing both at once: a mass consumer surface in ChatGPT and Codex, and a developer surface with two quality tiers, markup and SVG.
That matters because image generation is no longer a one-shot novelty. The competitive edge is shifting to edit loops, brand consistency and the ability to hand a result to another tool without re-prompting from scratch. GIF and SVG support push OpenAI further into that workflow layer. At the same time, speed claims of “50% faster” and “six seconds” are hard for outsiders to check without a fixed prompt set, so readers should treat them as OpenAI’s benchmarks until third-party boards catch up — the same caution TSN applies when covering other AI creative tool launches.
For UK and European users the product is already listed as available globally in ChatGPT and Codex. What still needs watching is how Flare and Sunburst behave under OpenAI’s existing content rules, and whether the API complexity dial becomes a real cost control or mostly a quality tweak. The next useful signal will be independent latency and edit-consistency tests, not another round of partner quotes.
Sources
Related stories
- Microsoft’s MAI Image 2.7 Preview Ranks in Top Three for Image Editing
- AI Creative Tools: Cheaper Video, Brand Kits and Claude Motion
- Canva Releases Craft, Its First In-House Image Generation Model
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
