Gemma 3n E2B

Overview

Google's Gemma 3n with 2B effective parameters (6B raw) using MatFormer, covering 140+ languages. Optimized for mobile and edge; text-only on Ollama.

When to pick this model

• Mobile and embedded deployments where memory is scarce
• Multilingual edge inference across 140+ languages
• Battery-constrained on-device chat
• MatFormer-based research and experimentation

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	2 GB
Q5_K_M	2.5 GB
Q8_0	3.5 GB
FP16 (no quantization)	6 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Strengths

• Built specifically for mobile and edge hardware
• 140+ language coverage in a tiny footprint
• MatFormer architecture maximizes memory efficiency
• Per-layer shared embeddings cut RAM use

Limitations

• 32k context only
• Absolute quality trails Gemma 3 9B
• Gemma license — not as permissive as Apache 2.0
• Multimodal features not exposed via Ollama

Architecture & training

Architecture: Gemma 3n E2B · on-device architecture · 2B effective · matPow

Training: Google Gemma 3n, optimized for mobile/edge with shared per-layer embeddings.

Quick start

ollama run gemma3n:e2b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.