RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can granite 4.0 h small run on the Mac M4 16GB?

No. Even at 2-bit with 8K context, granite 4.0 h small needs about 11.7 GB, more than the 10.7 GB usable on the Mac M4 16GB.

This is a mixture-of-experts model. llama.cpp can keep some expert layers in system RAM (--n-cpu-moe), so it may still run on this card with enough system memory, more slowly. This page does not model CPU or expert offload.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •11.6 GB No11.7 GB No12.1 GB No13.6 GB No—
3-bit (Q3_K_M) •15.0 GB No15.0 GB No15.4 GB No16.9 GB No—
4-bit (Q4_K_M) •18.8 GB No18.9 GB No19.2 GB No20.7 GB No—
5-bit (Q5_K_M) •21.9 GB No22.0 GB No22.4 GB No23.9 GB No—
6-bit (Q6_K) •25.3 GB No25.4 GB No25.7 GB No27.2 GB No—
8-bit (Q8_0) •32.6 GB No32.6 GB No33.0 GB No34.5 GB No—
16-bit (F16/BF16) •60.7 GB No60.7 GB No61.1 GB No62.6 GB No—

• real GGUF file size from ibm-granite/granite-4.0-h-small-GGUF. Speed is a bandwidth ceiling (120 GB/s), not a benchmark. Methodology.

Hardware that does fit it

RTX 4060 Ti 16GBRTX 4070 Ti SuperRTX 4080RTX 4080 SuperRTX 5060 Ti 16GBRTX 5070 Ti

More

granite 4.0 h small on every GPU and Mac · Every model that runs on the Mac M4 16GB · Apple Silicon guide · Try other settings in the calculator

Data: Hugging Face Hub (ibm-granite/granite-4.0-h-small), Apple specs. Model data last checked 2026-11-17; refreshed hourly.