GoogleGemma 3n

Gemma 3n E2B

Gemma 3n E2B uses selective activation for an effective 2B parameter size.

What the release provides

It accepts text, images, video and audio and returns text. The shared context is 32K tokens. Effective parameter size is smaller than the complete checkpoint.Google Gemma 3n model card

How to use it well

Practical guidance. Review the complete mobile workflow, including audio capture and image preparation. Measure sustained operation as well as a short demonstration. Keep unsupported actions outside the model's output contract and show the actual result returned by the application.

Specifications & access

Gemma 3n E2B specifications
DetailValue and conditions
AccessDownloadable
Model identifiersgoogle/gemma-3n-E2B-it