Gemma 4 E4B: A Local Agentic Edge Model in 6GB RAM

Melvin Vivas · X video post · 2026-04-04 · Open on X

Topics: AI Agents, Tool Use & MCP, LLM Fundamentals, Industry Trends & Job Market · Level: intermediate

Summary

Google DeepMind's Gemma 4 E4B, a small edge model quantized to 4-bit, completed a full repo audit by running Bash commands and tool calls locally. It needed only about 6GB of RAM, which shows that small open models can now run agent workflows on-device.

Key points

• Gemma 4 E4B is an edge-sized model from Google DeepMind.
• At 4-bit quantization it runs in about 6GB of RAM.
• It handled an agent task end to end: a full repository audit.
• The task used local Bash execution and tool calls, with no cloud API.
• Small quantized models are becoming practical for local coding agents.

Resources mentioned

• Gemma 4 · tool · ai.google.dev · free
  Google's family of open-weight models in several sizes, built to run on devices and offline, with multimodal and agentic abilities, and open to fine-tuning.
  Also in: Gemma 4 Runs Locally On-Device in the Antigravity SDK (Melvin Vivas on X · notes), On-Device AI: Running Gemma 4 E2B Offline on an iPhone with LiteRT (Melvin Vivas on X · notes), Running Gemma 4 Models Offline on an iPhone (Melvin Vivas on X · notes), Fine-tuning Gemma4-E2B on your own tweet style with Unsloth (Melvin Vivas on X · notes) and 25 more
• Google DeepMind (@GoogleDeepMind) on X · person · x.com · free
  Google DeepMind's official X account, which posts research and model releases such as Gemma.
  Also in: Running AI Locally: Rebuilding AIBackends as a Python Library with Open Models (Melvin Vivas on X · notes), Gemma 4 and Google DeepMind's Open-Source Push (Melvin Vivas on X · notes)

Try this

• Run Gemma 4 E4B (4-bit) locally and test it on agent tasks that use tool calls.
• Build a local repo-audit agent that runs a small quantized model and executes Bash tool calls.