OpenAI has added motion to ChatGPT's image tools. ChatGPT Images 2.6, released on November 4–5, 2026, lets users turn a text prompt or a picture into an animated GIF, then adjust its speed, edit individual frames and share it on repeat. The update also adds reusable reference images, and gives developers two new image models in the API, GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst.
OpenAI says the release is available across ChatGPT and Codex on web, desktop and mobile, on all subscription tiers, and that it generates images about 50% faster than the previous version. Here is a clear breakdown of what is new, what the numbers mean and what remains unconfirmed.
Key Takeaways
- GIF creation: Generate animated GIFs from a prompt or an image, with controls for speed, frame editing and looping.
- Reference images: Save likenesses of people, pets and characters and reuse them without uploading again.
- Two new API models: Flare for speed and Sunburst for quality and edit consistency, with SVG export.
- Speed: OpenAI claims about 50% faster generation, and Flare can produce an image in as little as six seconds.
- Usage: OpenAI says people now create more than 4 billion images a week across ChatGPT and the API, up from 3 billion in September.
Images 2.6 at a Glance
| Detail | Info |
|---|---|
| Release | November 4–5, 2026 |
| Where it works | ChatGPT and Codex, on web, desktop and mobile |
| Plans | All ChatGPT tiers, according to OpenAI |
| New consumer features | GIF creation, reusable reference images, image markup |
| New API models | GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst |
| Speed claim | About 50% faster than the previous version |
| API pricing | Not disclosed |
What's New in ChatGPT
GIF creation from prompts or images
Users can describe an animation in plain English or upload a photo and ask ChatGPT to bring it to life. Once the GIF is generated, they can:
- Adjust the playback speed.
- Inspect and edit individual frames.
- Share the finished GIF as a looping animation.
OpenAI's examples point to reactions, memes, animated characters and brand mascots, and even a basic sketch can be turned into a moving character.
Reference images that stay saved
A new reference-image feature keeps saved likenesses of people, pets and characters available for future creations, so users do not need to re-upload them each time. The goal is consistency: the same character or style can appear across several images and GIFs.
Better consistency across edits
OpenAI says the model is better at keeping character features, product designs and brand elements stable when a user makes several changes in a row. It also points to more professional-looking output, including posters and product images. Image markup, which lets users point to specific areas that need changing, was already available in ChatGPT.
Two New API Models: Flare and Sunburst
| Feature | GPT-Image-2.6 Flare | GPT-Image-2.6 Sunburst |
|---|---|---|
| Positioning | Speed tier | Quality and consistency tier |
| Headline claim | Images in as little as six seconds, at the low quality setting | Higher-quality generation and greater consistency across edits |
| Extras | Aimed at quick iteration | Image markup and SVG export for use in websites, apps and print |
| Best suited to | High-volume or interactive uses where speed matters | Detailed work and multi-step edits |
The API also gains a complexity parameter, which controls how much planning the model does before it generates an image. A lower setting suits quick edits, while a higher one can help with detailed requests. OpenAI has not yet published pricing for either model.
The Numbers, in Context
| Figure | What was reported | Context |
|---|---|---|
| 4 billion images a week | Across ChatGPT and the API, up from 3 billion in September | A company figure that does not separate consumer use from API use |
| 50% faster | OpenAI's claim for Images 2.6 over the previous version | The workloads behind the claim were not broken out. The previous release also advertised speed gains of up to 50%, so the comparison point matters |
| Six seconds | Flare's fastest generation time | Applies at the low quality setting, so higher quality settings will take longer |
Who Gets It, and When
OpenAI says Images 2.6 is available today in ChatGPT and Codex on all tiers, across web, desktop and mobile. Early reports on the first day were more cautious. They described a gradual rollout and noted that some accounts did not yet show the GIF option. Users who cannot see the feature should expect it to appear as the rollout completes, and may need to update their app or refresh the page.
How It Fits OpenAI's Image Strategy
- Steady cadence. Images 2.6 follows the September release of Images 2.5, which promised sharper detail and more precise editing. In one comparison test, that model split six categories evenly with Google's Nano Banana 2, though it still made mistakes when rendering text.
- Motion without a video app. OpenAI shut down its standalone Sora video app in March. GIFs are far simpler than full video, but they bring a lightweight form of motion generation into the main ChatGPT experience.
- Two tiers for developers. Offering a speed model and a quality model lets developers choose between cost, latency and polish.
Why It Matters for App Developers and Marketers
This section is ASOWorld's editorial analysis, not a statement from OpenAI.
Faster, more consistent image and GIF generation changes how teams produce creative assets:
- Cheaper creative testing. Short animated assets for social posts, ads and onboarding can be made and iterated in minutes, which suits A/B testing of messaging and visuals.
- Brand consistency at scale. Reusable reference images make it easier to keep a mascot or product look consistent across many assets, though final outputs still need human review.
- New in-app features. Sticker, avatar, messaging and creator apps can use the API to generate personalised images or animations, with the speed tier for interactive use and the quality tier for exports.
- SVG for design workflows. Vector export could help with icons and interface graphics that need to scale cleanly.
- Authenticity and rights. As AI-generated animations become common, teams should check platform rules, disclose AI use where required and avoid using real people's likenesses without permission.
For more coverage of AI platforms and their impact on app businesses, browse the AI Tech section on Marketing Trending.
📱 App Marketing Angle
Generative creative tools shorten the time between idea and asset, so the advantage shifts to teams that test quickly and measure what converts. Pair faster creative production with well-optimised store listings, clear positioning and strong ratings to turn that speed into installs. See how ASOWorld's AI app marketing solutions help AI apps turn new capabilities into rankings, installs and revenue.
What to Watch Next
- API pricing and limits. Costs and rate limits will decide how widely Flare and Sunburst are adopted.
- Independent comparisons. Tests against rival image models will show whether the speed and consistency claims hold up.
- GIF feature details. Length, quality and sharing options as the rollout completes.
- OpenAI's next releases. Further updates to its text and image models may follow in quick succession.
Quick Answers: ChatGPT Images 2.6
What is new in ChatGPT Images 2.6?
GIF creation from prompts or images, reusable reference images, better consistency across edits, and two new API models, Flare and Sunburst.
Can I make GIFs in ChatGPT now?
OpenAI says Images 2.6 is available on all tiers across web, desktop and mobile, though early reports suggested some accounts were still waiting for the GIF option.
What can I edit in a GIF?
You can adjust the playback speed, edit individual frames and share the GIF on repeat.
What is the difference between Flare and Sunburst?
Flare is the speed model, producing images in as little as six seconds at low quality. Sunburst focuses on higher-quality generation, edit consistency, image markup and SVG export.
How much do the new API models cost?
OpenAI has not disclosed pricing for either model.
How fast is Images 2.6?
OpenAI claims about 50% faster generation than the previous version, without specifying which workloads that applies to.
Follow Us on Facebook
