Google DeepMind·Gemma

Gemma 3n E4B

Used MatFormer and Per-Layer Embeddings to bring multimodal understanding to mobile devices.

LegacyLanguageMultimodalOpen weights

MODEL NUMBER104

Specifications

Only published specifications are shown.

Total parameters

Active parameters

Architecture

Hybrid

Context

32K

Input modalities

Text · Image · Audio · Video

Output modalities

Text

Open weights

Yes

License

Gemma Terms

Status

Legacy

Used MatFormer and Per-Layer Embeddings to bring multimodal understanding to mobile devices.

PreviousGemma 3 27B27B · 128K ctx

CurrentGemma 3n E4B8B · 32K ctx

New capabilities On-device Native audio

On-deviceNative audioVision

2025-06-26Open weights

Gemma 3n fully released its edge multimodal models.

Select a card to view the related model.