Image Model Releases Drive Surge in AI App Downloads, Revenue Gains Vary
Key Points
- Image‑model releases generate up to 6.5 × more downloads than traditional app updates, according to Appfigures.
- OpenAI's GPT‑4o image model added over 12 million installs in the 28 days after its launch.
- Google's Gemini Nano Banana contributed more than 22 million downloads, a four‑fold increase over prior releases.
- Meta AI's Vibes video‑feed drove 2.6 million extra installs but did not produce notable revenue.
- Revenue from image‑model spikes varies: GPT‑4o generated about $70 million, Nano Banana only $181 000.
- DeepSeek's R1 saw a 28 million‑download surge driven by industry curiosity rather than a visual model.
- The report warns that higher install numbers do not automatically translate into higher mobile earnings.
A new Appfigures report shows that releasing image‑generation models has become the most effective way for AI mobile apps to attract users, delivering up to 6.5 times more downloads than traditional updates. OpenAI's GPT‑4o image model, Google’s Gemini Nano Banana, and Meta AI’s Vibes each sparked multi‑million install spikes, though only OpenAI turned the surge into significant consumer spending. The findings suggest visual capabilities now trump chat‑only upgrades in driving app growth, but the revenue impact remains uneven across providers.
App intelligence firm Appfigures says AI developers are seeing their biggest download lifts when they launch new image‑generation models, not when they roll out conversational upgrades. The data, covering the last two years of major AI app releases, shows image‑model updates generate roughly 6.5 times more installs than traditional feature updates.
OpenAI’s GPT‑4o image model, released in March 2023, delivered more than 12 million incremental installs in the 28 days after launch. That figure represents a 4.5‑fold increase over the download spikes seen for the same company’s prior GPT‑4o, GPT‑4.5 and GPT‑5 releases, which were focused on text‑only capabilities.
Google’s Gemini followed a similar trajectory. The company’s Nano Banana image model, introduced alongside the Gemini 2.5 Flash release last August, added over 22 million downloads in the subsequent four‑week window—more than a four‑fold lift compared with earlier Gemini updates.
Meta AI entered the race with its Vibes video‑feed feature, technically a video model but still centred on visual content. Vibes generated an estimated 2.6 million extra installs after its September 2025 debut. While the numbers are smaller than OpenAI’s and Google’s, they still underscore a clear pattern: visual AI features draw users.
Despite the download boom, the revenue picture is less uniform. Nano Banana’s surge translated to roughly $181,000 in estimated gross consumer spending during its 28‑day post‑launch window—modest compared with the download surge. By contrast, OpenAI’s GPT‑4o image model produced an estimated $70 million in gross consumer spending over the same period, turning user curiosity into substantial revenue.
Meta’s Vibes, while boosting installs, did not generate any meaningful revenue in the tracked timeframe. The report cautions that higher install counts do not automatically equate to higher mobile earnings, highlighting the difference between user acquisition and monetization.
DeepSeek’s R1 model presents an outlier. The startup’s release in January 2025 sparked 28 million downloads, but the spike was driven more by industry buzz about its low‑cost training methods than by a new image‑generation capability. The case illustrates that curiosity and media coverage can also fuel install spikes, even when the underlying feature isn’t directly visual.
Overall, the Appfigures analysis signals a shift in how AI companies prioritize product updates. Visual capabilities now appear to be the primary hook for mobile users, eclipsing the earlier era where new conversational models like ChatGPT and Gemini’s text‑only upgrades sparked the biggest demand.
Developers looking to grow their user base may need to double‑down on image‑generation improvements, but they should also temper expectations about revenue. Converting the influx of curious downloaders into paying subscribers remains a challenge, as the data shows mixed outcomes across the major players.