Can Mac M4 Pro 24G run Gemma 4 31B IT?
Tight: GGUF Q3_K_M needs 18.3 GB at 4K, batch 1; headroom -0.3 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
| Quant | Total GB | Verdict | Weight sizing |
|---|---|---|---|
| GGUF Q4_K_M | 22.1 | Does not fit | Verified file |
| GGUF Q4_0 | 20.1 | Does not fit | Verified file |
| GGUF Q8_0 | 35.7 | Does not fit | Verified file |
| GGUF Q3_K_M | 18.3 | Tight | Verified file |
| GGUF Q5_K_M | 25.3 | Does not fit | Verified file |
Context capacity · GGUF Q3_K_M
| Context | Total GB | Verdict |
|---|---|---|
| 4K | 18.3 | Tight |
| 16K | 19.3 | Does not fit |
| 32K | 20.7 | Does not fit |
| 128K | 28.9 | Does not fit |
Generation speed
GGUF Q3_K_M: 12–16 tok/s · Derived range