v2.2 · 2026-06-12

VISION / MULTIMODAL MODEL — RUNTIME REQUIREMENTS

·Ollama's catalog page currently lists Text input for gemma3n tags, while Google's official documentation describes text, image, audio, and video inputs; treat local multimodal support as runner-dependent.

·RAM and VRAM figures are estimates based on GGUF file size plus conservative runtime and KV-cache overhead, not official hardware requirements.

Gemma 3n E4B Instruct

gemma-3n-e4b-it

MULTIMODAL

Gemma 3n E4B Instruct is a 8B multimodal model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.

PROVIDER

Google

FAMILY

Gemma 3n

MODEL TYPE

multimodal

PARAMETERS

8B (4B active)

MODALITIES

text, image, audio, video

ARCHITECTURE

MatFormer decoder architecture with Per-Layer Embedding caching, conditional parameter loading, MobileNet-V5 vision encoder, and USM-based audio encoder

CONTEXT WINDOW

32,768 tokens

TRAINING TOKENS

—

RELEASE DATE

2025-06-26

LICENSE & LINKS

Gemma Terms of Uselicense text

MODEL CARD

ai.google.dev/gemma/docs/gemma-3n

HUGGING FACE

google/gemma-3n-E4B-it

GGUF REPOSITORIES

bartowski/google_gemma-3n-E4B-it-GGUF community conversion

Community GGUF quantization of the official Google E4B instruction-tuned checkpoint.

lmstudio-community/gemma-3n-E4B-it-text-GGUF community conversion

LM Studio community text-only GGUF packaging; multimodal GGUF support may vary by runner.

SETUP HINTS

OLLAMA

Use ollama run gemma3n:e4b for the higher-capability effective 4B model; use gemma3n:e2b only when memory or latency is tighter.

LM STUDIO

Search for bartowski/google_gemma-3n-E4B-it-GGUF or lmstudio-community/gemma-3n-E4B-it-text-GGUF; verify whether the selected build supports multimodal input.

LLAMA.CPP

Use hf.co/bartowski/google_gemma-3n-E4B-it-GGUF:Q4_K_M for text inference; verify multimodal/audio support in the llama.cpp build before enabling non-text inputs.

QUANTIZATION FILE SIZES (GGUF)

Q4_K_M

4.24 GB

Q5_K_M

4.95 GB

Q8_0

7.35 GB

Quantization note: Q4_K_M for balanced local use; Q5_K_M when memory allows; use text-only GGUF if the target runner lacks multimodal support.

RAM / VRAM ESTIMATES

MIN RAM

8 GB

COMFORTABLE RAM

16 GB

MIN VRAM

6 GB

COMFORTABLE VRAM

8 GB

These are conservative local-inference estimates, not official hardware requirements.

HARDWARE FIT

CPU only (no GPU)

yes

8 GB RAM

yes

16 GB RAM

yes

32 GB RAM

yes

8 GB VRAM

yes

12 GB VRAM

yes

24 GB VRAM

yes

Apple Silicon (unified memory)

yes

Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.

BEST FOR

·mobile and laptop local assistants

·multimodal extraction

·speech and audio understanding

·efficient on-device chat

AVOID IF

·the local runner needs mature fully verified multimodal GGUF support

·a precise public training-token count is required

CAVEATS

·Selected E4B rather than E2B because E4B contains the E2B sub-model and is the better-supported higher-capability local choice while remaining small-device oriented.

·Gemma 3n uses effective-parameter terminology: E4B is an effective 4B configuration inside a larger raw-parameter model.

·Training-token count was not found in official Gemma 3n overview or release material, so it is null.

·RAM and VRAM figures are estimates based on GGUF file size plus conservative runtime and KV-cache overhead, not official hardware requirements.

SOURCE URLS