Phi-3-mini-instruct

A 3.8B parameter instruction-tuned language model from Microsoft with an MIT licence; in 4-bit GGUF format it uses approximately 4GB RAM, making it the recommended starting model for 16GB M2 Mac Mini deployments.

Details

• Services: text generation, instruction following

Related

gguf-4-bit-quantisation local-llm-inference llama-cpp ollama mac-mini-m2