Google has released Gemma 4 for smartphones through its AI Edge Gallery app, giving users access to an offline AI assistant that runs directly on their device without an internet connection, subscription or account.
The app is available on both iPhone and Android and supports text, image and audio tasks. Users can download Gemma models to their phone and run them locally, meaning prompts, conversations and files stay on the device rather than being sent to cloud servers.
Google AI Edge Gallery currently supports Gemma E2B and Gemma E4B, which are designed for smaller devices with lower power requirements. E2B is aimed at phones with less memory, while E4B is intended for higher-end devices with at least 8GB of RAM. Users can install the app, select a model and begin using it without creating an account or connecting to a remote service.
The launch comes as competition in AI becomes increasingly tied to subscription plans and premium tiers. Many of the leading consumer AI products now charge monthly fees, particularly for access to advanced models and faster response times.
Google appears to be taking a different route with Gemma 4. The company has positioned the models as open source under an Apache 2.0 licence, allowing developers to use, modify and build on them freely. Google says Gemma has already generated hundreds of millions of downloads and more than 100,000 community-built variants since the first release.
The app includes features beyond basic chat. Users can ask questions about images, transcribe audio, use prompt templates and access agent-style tools that handle tasks with help from maps, Wikipedia or community-built skills loaded from GitHub. Some versions also include a “Thinking Mode” that shows how the model works through a problem.
There are still limitations. Offline models generally perform best on newer phones with more storage and memory. Larger models can take up several gigabytes of space and may run more slowly on older devices. Community users on Reddit have reported that E2B works on mid-range phones, while E4B tends to perform better on newer flagship models. Some users have also pointed out that support for dedicated AI hardware acceleration is still limited in parts of the app.
Even with those trade-offs, Gemma 4 points to a wider shift in how AI may be used in the future. Instead of relying entirely on cloud-based systems, companies are increasingly pushing smaller, more efficient models onto personal devices.
For users concerned about privacy, data use or internet access, that could make local AI a far more practical option than it was a year ago.
Dear Reader,
Ledger Life is an independent platform dedicated to covering the Internet Computer (ICP) ecosystem and beyond. We focus on real stories, builder updates, project launches, and the quiet innovations that often get missed.
We’re not backed by sponsors. We rely on readers like you.
If you find value in what we publish—whether it’s deep dives into dApps, explainers on decentralised tech, or just keeping track of what’s moving in Web3—please consider making a donation. It helps us cover costs, stay consistent, and remain truly independent.
Your support goes a long way.
🧠 ICP Principal: ins6i-d53ug-zxmgh-qvum3-r3pvl-ufcvu-bdyon-ovzdy-d26k3-lgq2v-3qe
🧾 ICP Address: f8deb966878f8b83204b251d5d799e0345ea72b8e62e8cf9da8d8830e1b3b05f
Every contribution helps keep the lights on, the stories flowing, and the crypto clutter out.
Thank you for reading, sharing, and being part of this experiment in decentralised media.
—Team Ledger Life
Community Discussion