Gemma 4 26B A4B

Hot

Google's Gemma 4 MoE model — 26B total parameters with only 4B active, runs almost as fast as a 4B model. #6 open model on Arena AI

Model Specifications

ArchitectureVISION

Parameters26B

Familygemma

VRAM (Q4)13.0GB

Mixture of ExpertsActive inference parameters: 4B.

MoE architecture. 256K context. LMArena score 1441. Apache 2.0

#google#gemma#moe#multimodal#reasoning#flagship#trendingSource

Send this model's specs directly to your community.

Google's best open model

Efficient mid-size

Lightweight deployment

A complete breakdown of quantization levels and VRAM overhead for running local models.

Comparing NVIDIA and AMD options for the best speed-to-dollar ratio.

Understanding local AI formats and which one to pick for your specific hardware.