Image Generation

Gemini Can Now Make GIFs: Three-Second Loops in Eight Countries, Behind a Google AI Plan

November 21, 20262 views阅读中文原文
Gemini Can Now Make GIFs: Three-Second Loops in Eight Countries, Behind a Google AI Plan

The short version

On November 20, 2026, Google announced on the official Gemini account that the Gemini app can now generate GIFs. Four things worth knowing up front:

  • One prompt in, one animated file out. Type a description in the main chat box, or pick GIF from the image and video tools, and wait. What comes back is a looping animation of roughly three seconds.
  • The engine is a video model, not a new image model. It runs on Gemini Omni 1.2 Flash, which shipped on November 5. This release repackages it into a three-second consumer format.
  • Eight countries first. Australia, Canada, France, Germany, Japan, Korea, the UK, and the US. One English outlet listed only seven and left out France; Google's own help documentation includes it.
  • "Free" does not mean "no plan required." Google says the feature is available to Google AI subscribers at no extra charge in those countries. Personal accounts still need a Google AI plan, and work or school accounts need the matching Workspace license.

The detail most coverage gets wrong: this is a feature launch, not a model launch. The model has been in the developer API for weeks. What changed is that ordinary users can now reach it from a chat box.

This article is compiled from multiple media reports, together with Google's official help documentation and the company's official X account.

What this actually is

Gemini already had two separate visual pipelines. The Nano Banana family handles still-image generation and editing — swapping backgrounds, trying hairstyles, building stickers. The Omni family handles video. GIF generation plugs the second one into the chat interface.

In practice it works like any other image request. Open Gemini, open the image entry point in the left sidebar, choose a template or write a prompt. You can ask for a looping photo effect or a reaction meme. The only difference is that the output moves instead of sitting still.

There are two ways to take the result with you:

  • Copy and paste. Google says pasting into documents, email, messages, and social platforms preserves the animation in many applications.
  • Download the file. Save it locally as a .gif and send it to anyone; the recipient does not need Gemini.

You can also upload reference images, and you can use an avatar — a likeness generated once from your face and voice — as source material. That matters more than it sounds for anyone producing a series: you no longer have to redescribe "the same character" on every run.

What is genuinely new, and what is not

Only one thing is new: the entry point. Before this, producing motion with Omni meant going through the video generation flow with its native 15-second clips, or calling the API directly. Now there is an explicit GIF option. The trade is length — three seconds instead of fifteen — for something light enough to drop straight into a chat thread or a document.

The unchanged parts deserve more attention than the change:

  • Omni 1.2 Flash was released on November 5. Compared with the previous generation it extended native clips from 10 seconds to 15, and its developer API lists 720p output at $0.07 per second, roughly 30 percent below Omni 1.1. Those are video API figures, not a GIF pricing schedule.
  • The 15-second video inside the Gemini app is an Ultra-tier feature. Do not merge "15-second video" and "3-second GIF" into one story; they sit behind different entry points and different quotas.
  • Still images remain Nano Banana's job. Google's help documentation separates the two engines cleanly: images go to Nano Banana, motion goes to Omni.

A more accurate description: Google carved a small slice out of a capability it already had and turned it into a button in a chat box.

How to use it

  1. Open gemini.google.com, or the app, and sign in with a personal Google account that has a Google AI plan.
  2. Open the image entry point in the left sidebar, or just type the request into the chat box.
  3. Be specific in the prompt. "A looping GIF of a dog riding a surfboard" behaves far more predictably than "make something fun."
  4. If you need a consistent subject, upload a reference image first, or select your avatar.
  5. Copy and paste the result, or download the .gif.

One caveat: Google says the capability is rolling out gradually. Two people in the same country on the same plan can have different experiences — one sees the option, the other does not yet.

Good fit: who should try it now

  • Social and community managers. Reaction images, memes, seasonal motion graphics. This used to mean searching Giphy or Tenor for something close enough, or opening an editor. Now one description produces something nobody else has.
  • E-commerce and product pages. Turn a hero shot into a subtle looping light pass or transition; it holds attention better than a still, and three seconds loops cleanly without slowing a page.
  • Presentations, course material, internal docs. Paste it into a document, an email, or a message and the recipient needs no software at all — a lower barrier than sending a video link.
  • Series content that needs a consistent character. Avatar plus reference images keeps the same figure recognizable across multiple animations.

Not a good fit: who should wait

  • Anyone outside those eight countries. The help documentation is blunt: to create GIFs you must be in Australia, Canada, France, Germany, Japan, Korea, the UK, or the US.
  • Personal accounts without a Google AI plan. GIFs are generated by Gemini Omni, and Google states plainly that this requires a Google AI plan. A free Gemini account can still make still images with Nano Banana, but not this.
  • Work or school accounts without the right license. These are governed by Workspace licensing, not by simply having an account.
  • Users under 18. Image editing is documented as unavailable under 18, and media coverage treats GIF and video generation the same way.
  • Anyone who needs exact duration, frame rate, or resolution. Google has not published the GIF resolution, frame rate, or per-run limits. If you deliver assets to a spec sheet, this will not give you a guarantee.
  • Anyone who needs longer clips. Three seconds is three seconds. For product demos, talking heads, or narrative, use the video path.

Three common misreadings

First, "free" gets read as "no requirements." The wording is that the feature is free to Google AI subscribers in those countries — not free to all Gemini users. Merge those two claims and a lot of people will open the app and find nothing.

Second, the country lists do not match. The Verge and IT之家 both list eight countries including France. Android Authority lists seven and omits France. Google's help documentation includes France. Trust the documentation.

Third, "feature launch" becomes "model launch." Omni 1.2 Flash dates to November 5. Calling it a new model this week is simply wrong.

There is also a boundary most coverage skipped: video upload for editing is unavailable in the European Economic Area, Switzerland, the UK, and some US states. That does not block text-to-GIF, but if your plan was "turn this existing clip into a GIF," that route is closed in those regions.

Cost and time

  • Money: No additional charge in those eight countries, assuming a Google AI plan. There is no published per-image or per-clip price for GIFs.
  • Time: Google says generation can take a few minutes. That is not the one-second turnaround people expect from still-image tools, and first-time users tend to assume it has stalled.
  • Quota: Google has not published a separate GIF cap. It runs against your existing Google AI plan allowance.

What to do next

  • First, check whether you are in the rollout. Open Gemini and look for a GIF option in the image entry point. If it is there, you have it. If not, wait — it is not an account problem.
  • Second, test it against a real need. Skip "make something fun" and use an actual job: a product shot with a light sweep, a line of dialogue as a reaction image, a holiday greeting.
  • Third, verify two things. Does the animation survive pasting into the channel you actually publish on, and is three seconds long enough. Those two answers decide whether it belongs in your workflow.
  • Fourth, go back to the video model or the API when you need length or exact specs. GIF is a lightweight entry point, not a replacement.

The summary: this update moves an existing capability into a shallower entry point. Its value is convenience, not a new capability frontier. For content people, it saves the minutes spent hunting for a GIF and opening an editor. For anyone delivering to a spec, it is not yet ready to be written into a process.

Sources

  • Google Gemini Apps Help: https://support.google.com/gemini/answer/14286560
  • Google Gemini on X: https://x.com/GeminiApp/status/2123813017855103449
  • The Verge: https://www.theverge.com/tech/1038771/googles-gemini-app-can-now-make-gifs
  • Android Authority: https://www.androidauthority.com/gemini-gif-generation-rollout-3737901/
  • IT之家: https://www.ithome.com/1/030/381.htm
  • Magica (context and limits): https://magica.com/news/gemini-gif-generation-paid-access

Sources

Comments

0 published
Log inorsign up to leave a comment.

Loading comments...