Image AI model releases are generating 6.5 times more app downloads than traditional model updates, Appfigures found. The firm measured download spikes in the 28 days after new model launches and saw image-focused releases lift installs far more than chatbot-only upgrades. Google’s Gemini 2.5 Flash image model and OpenAI’s GPT-4o image model each produced double-digit millions of extra downloads. But Appfigures cautioned that more installs didn't always translate into higher mobile spending.

Appfigures tracked installs around major AI model releases and found a clear shift in what draws users. Image-capable models now trigger the biggest surges. Appfigures compared downloads in the four weeks after a model launch to the app’s baseline. The result: image model drops produced 6.5 times more downloads than traditional conversational model updates.

That change reverses an earlier pattern. New conversational features used to drive big interest. Voice chat and smarter text replies once brought tens of millions of installs. Now the visual features are doing the heavy lifting. People are installing apps to try image generation, editing, and video features they couldn't use before.

Large spikes tied to specific image releases

Two headline examples show the scale. For Google’s Gemini, the firm released an image model named Nano Banana last August as part of the Gemini 2.5 Flash update. Appfigures counted more than 22 million extra downloads in the 28 days after that image-model rollout. Downloads for Gemini climbed to more than four times their usual level over the same window.

OpenAI’s ChatGPT saw a similar effect. After OpenAI added image capabilities to GPT-4o, ChatGPT picked up a substantial increase in installs in the following 28 days.

Mastheads

Appfigures said that gain was notably larger than the downloads the app saw after prior text-focused releases.

Other releases produced smaller but still noticeable bumps. Meta’s Vibes feature, a feed driven by visual and short-video models, added a smaller uplift in the 28 days after its September launch. Appfigures noted that Vibes is technically a video model, but it fits the broader visual-content category that's now pulling users in.

More downloads, mixed returns

Not all download spikes turned into money. Appfigures compared incremental installs with short-term consumer spending and found wide variation. The Nano Banana spike drove only modest estimated consumer spending in the 28-day window after launch; that sum was small relative to the download surge.

Some releases added downloads but produced little measurable revenue in that same window, Appfigures said. By contrast, at least one of the major image-model releases showed clearer signs of short-term monetisation; others had far lower short-term conversion.

The report’s takeaway is direct. Image models give users a reason to install and try an app’s visual features. They don't automatically create paying customers. Downloads spike. Converting those users into subscribers or buyers requires a separate step.

Developers shifting priorities

The data suggests mobile developers will push visual capabilities harder. Teams are now prioritising image generation, editing tools, and short-form video features in app updates. Those features can be easier to demo and share on social media than incremental chatbot improvements. People can see and post images quickly, which helps drive word of mouth and more installs.

At the same time, the numbers remind developers that user acquisition and monetisation are distinct challenges. A big install wave can be a testing ground. But retention, conversion rates, and pricing need work if the install bump is to turn into sustainable revenue.

Related Articles

Appfigures' 28-day comparison shows image-focused updates drive big install spikes, but turning those users into paying customers remains developers' bigger task.

This article was created with AI assistance.