by 
June 03, 2025 : Google is set to launch a new app called ‘AI Edge Gallery’ for Android users, enabling them to utilize artificial intelligence (AI) on their smartphones without an internet connection. This app allows users to generate images, write code, and get answers to various questions offline, ensuring both data privacy and eliminating the need to rely on server responses.
At the core of the ‘AI Edge Gallery’ app is Google’s proprietary language model, ‘Gemini 3-1B’, which is only 529 megabytes in size. Capable of processing 2,585 tokens per second, this model enables quick completion of tasks like text generation, document analysis, and providing smart replies in messaging apps.
The app is built on Google’s ‘AI Edge’ platform, leveraging technologies like TensorFlow Lite and MediaPipe to optimize AI performance based on the phone’s hardware. However, Google has cautioned that performance may vary depending on the device. Older or less powerful phones may face challenges running larger models, and lighter models are recommended in such cases.
The app features two primary functions—‘AI Chat’ and ‘Ask Image’—offering text and image-based AI assistance. Additionally, a ‘Prompt Lab’ section allows users to test model responses with single queries, supported by ready templates and response settings.
Currently in its experimental ‘Alpha’ phase, the app is released as open-source under the Apache 2.0 license, allowing users to use, modify, or employ it for commercial purposes. Google has also announced that an iOS version of the app will soon be available for iPhone users.
AI/MR
.jpeg)
