Can I run it?

Can I run Phi 4 mini instruct on Apple M5?

Runs

Runs fully in unified memory on your Apple M5 at BF16, 8K context, about 9.6–13 tok/s. Weights 7.15 GB plus 1.00 GB of KV cache — 8.52 GB of 11.0 GB usable. Up to 16K context still fits.

Best quantBF16

Working context8K

Max that fits16K

Speed9.6–13 tok/s

Weights7.15 GB

KV cache1.00 GB

Total8.52 GB

Usable11.0 GB

Phi 4 mini instruct details · Apple M5 page · How this estimate is made

Run commands

llama.cpp

llama-server -hf unsloth/Phi-4-mini-instruct-GGUF:BF16 -c 8192 -ngl 32

Ollama

ollama run phi

LM Studio

lms get unsloth/Phi-4-mini-instruct-GGUF
# Then load quant BF16

Assumptions

Estimate, not a benchmark. Decode speed uses a memory-bandwidth model at 55–75% of peak.
KV cache is FP16. Engine overhead assumes llama.cpp (0.3 GB + 1% of the weights). Ollama is closer to 0.9 GB + 3%, LM Studio closer to 1 GB + 3%.
BF16 size is the real GGUF file size from Hugging Face.
Unified memory uses 75% for the GPU by default, and always leaves about 5 GB for the OS.
Speed is quoted at 8K context, not at the maximum context that still fits.

Larger models that also fit

Can I run Qwen3.8 27B on Apple M5? — Tight fit, UD-Q2_K_XL, 8.1–11 tok/s
Can I run gemma 4 12b it on Apple M5? — Tight fit, Q6_K, 6.5–8.8 tok/s
Can I run Qwen3.5 9B on Apple M5? — Tight fit, Q8_0, 8.6–12 tok/s

Updated 2026-10-17. Profile on this page is the device default.