Model family

Gemma 4

Dense

Listed checkpoints

• mlx-community/gemma-4-e2b-it-4bit

Notes

Checkpoints listed here identify models in this family. Check each recipe for inference verification, hardware measurements, and requirements before running it.

Feature matrix

Runtimes	`generateserve`
Backends	CUDA:partially supportedMetal:supportedROCm:partially supported
Tensor parallel	yes
Pipeline parallel	yes
Drafters	`mtp`
KV modes	`fp16int8turbo4`
Input modalities	`text`
Output	tokens
Model types	`gemma4gemma4_text`

Recipes in this family

Choose hardware to estimate fit.

Google

Gemma 4 31B IT

Use for Chat

Model ID: google/gemma-4-31B-it

Dense · 31B · Context: 262,144 · CUDA, METAL

MLX Community

gemma 4 e2b it 4bit

Use for Chat

Model ID: mlx-community/gemma-4-e2b-it-4bit

Dense · Information unavailable · Context: 131,072 · CUDA, METAL