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 Model | huggingface:unsloth/gemma-4-12b-it |
| Engine | llama.cpp |
| Quantization | Q4_K_XL |
| Platform | metal |
| VRAM Required | 8GB 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...