Artificial intelligence (AI) image processing models are becoming the main driving force driving the growth of mobile applications, surpassing the previous role of chatbot upgrades. This is a noteworthy conclusion in a new report from Appfigures (a company providing data analysis platforms for mobile applications based in the US).
According to data, the release of new image creation models helps the number of AI application downloads increase 6.5 times compared to traditional model upgrades.
This trend shows that users are increasingly attracted to the ability to create visual content instead of just text interaction or conversation.
Previously, advances in chatbots such as improving conversation capabilities or adding voice interfaces were the main factors attracting new users. However, the context is changing rapidly as AI creating photos brings an intuitive, easy-to-share and more curious experience.
Specifically, both ChatGPT and Google Gemini recorded tens of millions of new downloads after launching image models.
With Gemini, the introduction of the Nano Banana image model and the Flash Gemini 2.5 version has helped the application increase by more than 22 million downloads in just 28 days, more than 4 times higher than before.
Meanwhile, ChatGPT reached more than 12 million installs after launching the GPT-4o image model, about 4.5 times higher than previous versions such as GPT-5 or GPT-5. These figures show the strong attraction of AI image in attracting new users.
A similar trend also appeared in Meta, when the Meta AI platform launched the AI video feature Vibes, helping to increase about 2.6 million downloads in 28 days. Although it is a video model, this product still focuses on image experience, contributing to strengthening the trend of prioritizing visual content.
However, the report also pointed out a paradox that the increase in downloads does not mean increased revenue. Many users download applications to try the photo creation feature but do not convert to paid users.
For example, Gemini's Nano Banana model only generated about 181,000 USD in revenue in 28 days, although the load increased sharply. Similarly, Meta AI recorded load growth but did not generate significant revenue.
Among the major platforms, only ChatGPT showed effective conversion capabilities. The GPT-4o image creation model helped OpenAI's consumer revenue increase by about 70 million USD in 28 days after its launch.
The report also mentions DeepSeek as a special case. Although the DeepSeek R1 model attracted up to 28 million downloads, this is considered a boom effect thanks to market curiosity about low-cost technology, rather than the AI image trend.
In general, data shows that AI is entering a new phase, where image experience becomes a key factor to attract users. However, the problem of transitioning from downloads to revenue is still a major challenge for technology companies.
