Image model releases are driving significant growth for AI mobile apps, generating 6.5 times more downloads than traditional model updates, according to a report from Appfigures. This shift indicates a new trend where releases focused on image capabilities lead to a surge in user engagement, contrasting with earlier periods when updates to conversational models were the primary drivers of demand.
ChatGPT and Google’s Gemini both saw substantial increases in downloads following their respective image model releases. Gemini’s image model, Nano Banana, contributed over 22 million downloads in the 28 days following the introduction of the Gemini 2.5 Flash image model last August, increasing downloads by more than four times during that period.
ChatGPT experienced similar success, adding over 12 million incremental installs within the 28 days following the launch of its GPT-4o image model in March 2023. This marked approximately 4.5 times the downloads compared to its previous model releases, including GPT-4o, GPT-4.5, and GPT-5.
Other models also mirrored these trends, although to a lesser extent. Meta AI’s video feed, Vibes, achieved around 2.6 million incremental downloads within the first 28 days after its September 2025 launch. Despite this growth, the report cautioned that increased downloads do not automatically translate into higher mobile revenue.
New image model releases attract users to test enhanced features but do not guarantee conversion to paying subscribers. For example, while Nano Banana prompted over 22 million downloads, it only generated an estimated $181,000 in gross consumer spending in the same time frame. Meta AI’s Vibes similarly garnered significant downloads but did not lead to meaningful revenue.
In contrast, ChatGPT successfully converted the spike in user interest into revenue. OpenAI’s GPT-4o image-generation model resulted in an estimated $70 million in gross consumer spending over the 28 days after its launch, surpassing its prior baseline.
Appfigures also analyzed DeepSeek’s R1 model, which garnered 28 million downloads following its January 2025 release. This case was noted as atypical; rather than a standard model comparison, DeepSeek’s launch marked its emergence as a notable player in the tech landscape, driven by innovative training techniques. This instance underlines that curiosity, rather than specific model features, can also significantly influence download figures.
