Back to registry
METAL8GB VRAM

arnav080/gemma-4-12b-Q4_K_XL-M116GB

Gemma 4 12B recipe for local execution using llama.cpp based on Unsloth guide

Configuration Specifications

Base Modelhuggingface:unsloth/gemma-4-12b-it
Enginellama.cpp
QuantizationQ4_K_XL
Platformmetal
VRAM Required8GB minimum

Telemetry Benchmarks

No runs recorded yet

Be the first to benchmark this recipe! Run the CLI command from your terminal:

bloc run arnav080/gemma-4-12b-Q4_K_XL-M116GB
gemma-4-12b-Q4_K_XL-M116GB.yaml
Loading workspace editor...