Google AI Edge Gallery
Highlights
- ▸Gemma 4 E2B (2.54GB) running natively on iPhone is fast enough to be useful — on-device LLMs crossed the 'real workload' threshold on consumer hardware
- ▸The 'skills' demo is the real signal: tool-calling against HTML widgets works with a phone-scale model, not just frontier models
- ▸Private, offline, zero-latency on-device inference opens use cases that cloud models structurally can't serve — field work, privacy-sensitive domains, latency-critical loops
Original excerpt
Sponsored by: Honeycomb — AI agents behave unpredictably. Get the context you need to debug what actually happened. Read the blog
Google AI Edge Gallery (via) Terrible name, really great app: this is Google's official app for running their Gemma 4 models (the E2B and E4B sizes, plus some members of the Gemma 3 family) directly on your iPhone.
It works _really_ well. The E2B model is a 2.54GB download and is both fast and genuinely useful.
The app also provides "ask questions about images" and audio transcription (up to 30s) with the two small Gemma 4 models, and has an interesting "skills" demo which demonstrates tool calling against eight different interactive widgets, each implemented as an…
Frequently asked questions
What is "Google AI Edge Gallery" about?
This article by Simon Willison is part of the Simon Willison reading list on Burn 451, covering llm blog.
Who wrote "Google AI Edge Gallery"?
This piece is part of the Simon Willison vault on Burn 451, covering llm blog. The original author is attributed at the source link.
How can I read more content from Simon Willison?
The complete Simon Willison reading list is available at burn451.cloud/vault/simon-willison. Each article includes an AI-generated summary so you can decide what to read in seconds. Connect the Burn 451 MCP server to Claude or Cursor to query all Simon Willison articles as live AI context.
Can I use "Google AI Edge Gallery" with Claude or Cursor?
Yes. Install the burn-mcp-server npm package and connect it to Claude Desktop, Claude Code, or Cursor. Once connected, your AI can search and reference this article and the full Simon Willison vault in real time — no manual copy-paste required.
16 more articles in this vault.
Import the full Simon Willison vault to Burn 451 and build your own knowledge base.
Content attributed to the original author. Burn 451 curates publicly available writing as a reading index. For removal requests, contact @hawking520.