AI news · OpenAI · Images 2.6
OpenAI’s Images 2.6 brings GIFs to ChatGPT at 2.5’s token price
The new image models list at the same API rate as September’s, but they sit outside OpenAI’s half-price Batch option for jobs that can wait.
Describe an image or upload a photo, and ChatGPT now returns an animated GIF in one step. In the same chat you can change the playback speed or edit single frames. The Verge reports that the Images 2.6 update of 4 November 2026 brings the feature, and that OpenAI is rolling it out to all ChatGPT and Codex users.
OpenAI says Images 2.6 generates faster, and puts weekly output at more than 4 billion images across ChatGPT and its API, up from 3 billion in September.
Developers get two API models, GPT-Image-2.6 Flare and Sunburst, and both can return animated GIFs and SVG, a vector format that stays sharp at any size (changelog). OpenAI’s model pages point Flare at interactive generation and editing and Sunburst at detailed compositions and edits.
Both list at $30 per million image output tokens, the units OpenAI bills image generation in. That is the rate of the 2.5 models released on 8 September. The change sits in the Batch column, OpenAI’s half-price route for jobs that can wait. GPT Image 2 and 2.5 appear there at $15. The 2.6 models do not.
OpenAI gives no per-image price for 2.6, only token rates and a calculator. For scale, our 4 November check has a 1K image at $0.0336 on Nano Banana 2.1 and $0.067 on Nano Banana 2. OpenAI’s medium GPT Image 2 square cost $0.053 in output tokens alone when we read it on 23 September, with prompt and input tokens billed on top (our ChatGPT page).
A catalogue run that can wait therefore bills at $15 per million output tokens on 2.5 through Batch against $30 on 2.6, which leaves 2.6 to earn its rate on work someone is waiting for, or on the GIFs and SVGs that came with it.
The prompt
Create an animated GIF, square 1:1, a short seamless loop for a social media ad. A slim matte teal drinks can with the brand name "NOVA" printed large in bold cream capitals on its front turns one slow full rotation on the spot, standing on a warm sand-coloured paper backdrop that fills the whole frame. Small cold water droplets sit on the can and catch the light. Soft daylight falls from the upper left, and one soft shadow rests on the paper to the right of the can, turning gently as the can turns. In the top quarter of the frame, the words "ICE COLD" stay fixed in bold cream capitals, about half the frame wide, with a clear band of empty backdrop above them and on both sides. Straight-on camera at can height, the can centred in the lower three quarters of the frame, with empty backdrop below it. The camera, the backdrop, the words and the light stay still; only the can and its shadow move. The last frame matches the first so the loop has no jump.Sources
- The Verge: ChatGPT will now make you animated GIFs. Checked
- OpenAI: Changelog. Checked
- OpenAI: Pricing. Checked
- OpenAI: GPT-Image-2.6 Flare Model. Checked
- OpenAI: GPT-Image-2.6 Sunburst Model. Checked
Alexandre S.
Alexandre S. is a creative ad strategist and copywriter with ten years' experience in digital marketing across multiple channels.
More AI news from
- Uber counted four AI-agent rides in a week of 320 million trips
- Google opens AI Mode checkout to Canadian shoppers
- Criteo offers ChatGPT ads in 89 countries as revenue falls 11%
- AppLovin revenue rises 55% as it rolls out new ad models
- Microsoft’s Clarity now shows what AI agents do on your site
- Trade Desk ties Koa to Dstillery’s agent for mid-flight audiences
- Anthropic strikes first news licensing deal, with AP, Adweek says
- Reddit trials ads in its own search as outside traffic falls
- Fal’s valuation doubles to $16bn, Bloomberg reports