Google Gemini Now Turns Text Prompts Into GIFs

Google has added GIF generation to its Gemini assistant, letting users turn a plain-language prompt into a short looping animation. The company announced the rollout in a post on X on Thursday 20 November, and its Gemini help page says the feature accepts text, images and videos as inputs, replacing a search through fixed reaction libraries with a result generated for the request.
The output is a silent loop of roughly three seconds. Users can copy, download or share the result in documents, email, messaging services and social apps that support GIFs. To make one, users open Gemini and start a chat, or select the image or video creation tool, choose GIF from the format selector, describe the scene and optionally attach an image or video as a reference. A still photo can become a subtle loop in the style of a cinemagraph, while reference media can help keep a character or design consistent across variations.
Eight countries, no subscription
The feature is initially available in eight markets: Australia, Canada, France, Germany, Japan, South Korea, the United Kingdom and the United States. It is free, with no Google AI subscription required, and the rollout is gradual, so eligible users may not see the GIF format immediately. Google has not said when the feature will reach other countries. The help page presents it as a consumer feature of the Gemini app, available through the main chat or the image and video tools.
Powered by Gemini Omni 1.2 Flash
Behind the feature sits Gemini Omni 1.2 Flash, Google's latest video generation model, according to the company's model documentation. The multimodal model supports video generation, editing, extension and upscaling, and its broader video output can include synchronised audio, run for up to 15 seconds and use resolutions from 360p to 4K. GIF generation uses the short, silent, looping portion of that capability. The model also accepts as many as ten reference images and three reference videos of up to 15 seconds each, although Google's consumer instructions do not indicate that every reference control is available through the GIF selector.
Following OpenAI into animated replies
The launch follows OpenAI's rollout of GIF generation in ChatGPT, which arrived a couple of weeks earlier using its Images 2.6 model. The two companies are now matching each other feature for feature in generative video and animation inside mainstream assistants. One boundary remains for developers: Google has presented the GIF control as a Gemini product feature and has not documented a dedicated GIF-generation endpoint, so anyone building GIFs into an application may still need to generate a clip, remove the audio, encode it as a GIF and configure the loop themselves. Google says its application programming interface price for 720p generation has fallen by 30 per cent to about 0.07 dollars per second, while the consumer GIF feature stays free in the supported markets.