Gemma 3n E4B

Available

Google DeepMindOpen weights

Google's mobile-first Gemma 3n model variant, built with a MatFormer-style architecture for efficient on-device multimodal inference. The E4B variant has roughly 4B effective parameters, supports text, vision, audio, and video-oriented use cases, and is released under Gemma terms.

Official page ↗Model card ↗

Specifications

License: Open weights · Gemma Terms of Use
Weights: Downloadable
Architecture: hybrid
Parameters: 8B · 4B active
Context window: 32K tokens
Max output: —
Knowledge cutoff: —
Price (in / out, $/M): —
Modalities: TextVisionAudioVideoCode

Benchmarks

No benchmark scores recorded yet. Spotted some? Submit a correction.

Vendor-reported figures are claims until independently verified. See methodology.