The rapid evolution of generative AI is reshaping product design across major tech companies, including Google. While many AI features currently rely on powerful remote servers, Google is preparing to empower developers with new mobile AI tools. At the upcoming I/O event, the tech giant is expected to unveil a suite of APIs that will allow developers to harness the capabilities of Gemini Nano for on-device AI applications.
New Developments in ML Kit
In a strategic move, Google has discreetly released documentation detailing significant new AI features for developers. As reported by Android Authority, an update to the ML Kit SDK will introduce API support for on-device generative AI functionalities through Gemini Nano. This innovative technology is built on AI Core, akin to the experimental Edge AI SDK, and integrates seamlessly with existing models, offering a set of predefined features that promise ease of implementation for developers.
According to Google, the ML Kit’s GenAI APIs will enable applications to perform tasks such as summarization, proofreading, rewriting, and image description—all without the need to transmit data to the cloud. However, it is important to note that Gemini Nano’s capabilities are not as extensive as those of its cloud-based counterpart, which may lead to certain limitations. For instance, summaries generated can only consist of a maximum of three bullet points, and image descriptions will be restricted to the English language. Additionally, the quality of outputs may fluctuate depending on the specific version of Gemini Nano utilized on a device.
The standard version, Gemini Nano XS, occupies approximately 100MB of storage, while the Gemini Nano XXS, as seen on the Pixel 9a, is a more compact alternative at just a quarter of that size. This smaller version is text-only and features a significantly reduced context window.
Credit: Ryan Whitwam
This initiative bodes well for the Android ecosystem, as ML Kit is compatible with devices beyond Google’s Pixel lineup. While Pixel devices will leverage Gemini Nano extensively, several other smartphones are already optimized for this model, including the OnePlus 13, Samsung Galaxy S25, and Xiaomi 15. As more devices incorporate support for Google’s AI model, developers will find ample opportunities to enhance their applications with generative AI features tailored to a broader audience.
