Gemini Nano: Discover the Android version of Google’s most powerful AI

This Wednesday (6) Google introduced Gemini Nano, a “portable” version of its new multimodal language model (LLM). This is the company’s most efficient artificial intelligence (AI) set to run natively on mobile phones.

Google explains that the Gemini Nano looks like a compact version of the standard Gemini. The model is prepared to rely on special cores for artificial intelligence tasks, even if the device is not connected.

According to the company, it includes mainstream capabilities such as text summarization, contextual responses, grammar correction, and advanced proofreading.

“For example, Gemini Nano’s advanced language understanding allows the Pixel 8 Pro to concisely summarize content in the Recorder app, even when the phone’s network connection is offline,” the company notes.

Gemini Nano on Android

On Android, Gemini Nano is accessed via the AICore service. This new part of the system manages the template, workflows, security features, and more.

Thanks to this service, developers will be able to benefit from the capabilities of local artificial intelligence. In AICore, programmers will be able to use Low Order Adaptation (LoRA) to enhance Gemini Nano and train the model for specific use cases.

Gemini Nano on Pixel 8 Pro

Where the Gemini Nano really shines right now is on the Pixel 8 Pro. The top level provides the basis for the model to perform in the best possible way and without the need for the Internet.

In the Pixel 8 Pro’s Recorder app, Gemini Nano can be used to create text summaries of audio conversations like interviews, discussions, presentations, and more.

Another interesting feature is Gboard’s “Smart Reply” feature. This feature suggests high-quality canned replies for WhatsApp conversations (more messengers will be added soon).

Smart Reply is currently available in developer trial. There are no predictions regarding the exact launch of the new product.