Skip to content

Can Mac M4 24G run Nanbeige4.3 33B?

Tight: GGUF Q3_K_M needs 17.6 GB at 4K, batch 1; headroom 0.4 GB.

Text input, FP16 KV cache, one card. Unified-memory capacity uses the GPU guide’s usable limit; negative headroom may require host offload. Speed source settings may differ from this 4K estimate.

Measured runs

No measured run for this pair. Sizing is an estimate, not a runtime test.

Quant memory at 4K

QuantTotal GBVerdictWeight sizing
GGUF Q3_K_M17.6TightVerified file
GGUF Q4_K_M22.6Does not fitVerified file
GGUF Q8_037.7Does not fitVerified file

Context capacity · GGUF Q3_K_M

ContextTotal GBVerdict
4K17.6Tight
16K20.2Does not fit
32K23.8Does not fit
128K45.3Does not fit

Generation speed

GGUF Q3_K_M: 8 tok/s · Bandwidth ceiling (not measured)

Other model / GPU pairs