Skip to content
Log in

google/gemma-4-26B-A4B-it

huggingface_int4_embedl-mtp_tensorrt-edgellm
google/gemma-4-26B-A4B-it architecture graph
google/gemma-4-26B-A4B-it architecture graph
  • huggingface
  • AGX Thor
Labels
  • Image
  • Text
  • VLM
  • Generation
Build
Embedl-optimized
Precision
W4A16
Runtime
TensorRT Edge-LLM 0.10.1 with Embedl optimizations
Context length
4,096 tokens

Google's mixture-of-experts Gemma 4: 26B parameters, about 4B active per token, with images.

Files

Image Size

Model

docker.io/embedl/model-artifacts:gemma4-26b-a4b.huggingface_int4_embedl-mtp_tensorrt-edgellm.d89a196778b10a14

 17.83 GiB

Runtime

docker.io/embedl/models:runtimes.tensorrt-edgellm.arm64.2026101815014ee0

 1.01 GiB