A report from Appfigures highlights a significant trend in the app market, revealing that the introduction of image models is markedly boosting downloads for AI mobile applications, generating 6.5 times more installations than traditional model updates. This shift contrasts with earlier patterns, where new conversational models and features led to increased demand.
For example, following the launch of their image models, both ChatGPT and Google’s Gemini witnessed a surge in downloads. Gemini’s image model, Nano Banana, added over 22 million downloads within 28 days post-launch, a fourfold increase in its download rate. Similarly, ChatGPT gained more than 12 million installs after unveiling its GPT-4o image model last March, representing 4.5 times more downloads than previous model releases.
Other models displayed similar trends, albeit at smaller scales. Meta’s AI video feed, Vibes, added approximately 2.6 million downloads after its September 2025 release. Though this was technically a video model, it ultimately fits into the broader category of visual content enhancement.
However, the report warns that spikes in downloads do not necessarily equate to increased revenue. While new image model launches entice users to try out enhanced image generation features, translating this interest into subscriptions remains challenging. For instance, Nano Banana drove an estimated $181,000 in gross consumer spending in its first month, despite its huge download increase, and Meta’s Vibes also contributed additional downloads but failed to generate significant revenue.
In contrast, ChatGPT effectively capitalised on this surge, with its 4o model generating an estimated $70 million in consumer spending within the same timeframe. This contrasts with OpenAI’s previous model launches, where revenue did not see a similar uptick despite comparable download increases.
An exception in this trend was seen with DeepSeek R1, which gained 28 million downloads after its January 2025 release. However, this surge was driven more by the technology’s innovative approach to training AI models rather than the launch of a traditional image model, illustrating how curiosity and novelty can attract interest.
In summary, while the launch of image models is driving substantial download increases for AI apps, translating these metrics into significant revenue remains a complex challenge for developers.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
