Apple · 64 GB LPDDR5X · 273 GB/s bandwidth ·
DeepReinforce · 35.0B (active 3.0B) · MoE · model ctx up to 256K
Yes — the Mac mini M4 Pro 64GB (64 GB LPDDR5X) can run Ornith-1.0-35B. At Q4 with the default 8K context it needs ≈ 20.85 GB VRAM (weights 19.25 GB + KV cache 0.6 GB + 1 GB runtime overhead), which fits within 64 GB. Decode at Q4/8K: ≈ 57.9 tok/s (estimated, batch 1). FP16 needs ≈ 71.6 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology.
Fit & speed by quant and context
Quant
Context
VRAM needed
Fits?
tok/s (decode)
Q4
4K
20.55 GB
Yes
57.9est.
Q4
8K default
20.85 GB
Yes
57.9est.
Q4
32K
22.65 GB
Yes
57.9est.
FP16
4K
71.3 GB
No
36.7est.
FP16
8K default
71.6 GB
No
36.4est.
FP16
32K
73.4 GB
No
34.6est.
Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured
anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s
means the model is far beyond this card (offload-only territory).