Google has unveiled its latest innovation in large language model (LLM), Gemini with three distinct versions — Nano, Pro, and Ultra. Each of them has its own unique capabilities, and if you’re curious, you can read all about them in our separate story.

Google Pixel 8 Pro with Gemini Nano AI model

Gemini Nano is the smallest and most efficient AI model from Google. This model is designed to run locally on your phone and Pixel 8 Pro is the first smartphone to use it. And, since Gemini Nano runs locally on the smartphone, users need not be connected to the internet. 

Pixel 8 Pro features powered by Gemini Nano

This local processing (currently) powers two new features for the Pixel 8 Pro:

  • Summarize in Recorder: This feature lets you automatically generate bullet-point summaries of audio recordings, enhancing productivity and information absorption.
  • Smart Reply in Gboard: This upgrade suggests contextually relevant replies within messaging apps, but currently is only limited to WhatsApp. However, Google plans to expand this feature to other apps soon. As Google says, Smart Reply saves you time by offering high-quality responses that are both conversational and contextually aware.

These features are part of the December Feature Drop, which also includes many other new additions (we’ll cover them in a separate article).

While the Tensor G3 chip currently has the exclusive power to run these AI features, it’s currently limited to the Pixel 8 Pro. However, we expect it to become available on the standard Pixel 8 soon, too.

Likewise, Google is also working to bring Gemini Nano ability to the entire Android system. They have launched a new system called AICore to help developers integrate Gemini-powered features into their apps.

While your phone will still need a powerful processor, Google confirms compatibility with processors from Qualcomm, Samsung, and MediaTek. Developers can already join Google’s early access program.

Related:

(Source: 1, 2)

RELATED ARTICLESMORE FROM AUTHOR