Skip to content
Model library

Google DeepMind

Gemma 4 26B-A4B

A multimodal mixture-of-experts model for reasoning, coding, and assistants that understand images.

Language models
Parameters
26B total / 4B active
Input
Text / images / video
Output
Text
Context
Up to 256K tokens

Build with it

  • Multimodal assistants
  • Coding and reasoning
  • Image and document analysis
Original model card & license