OwnRig

Qwen 2.5 72B Instruct on Apple M4 Max (128GB Unified)

Apple M4 Max (128GB Unified) can run Qwen 2.5 72B Instruct at 6 tok/s at Q4_K_M, though performance is acceptable. Consider a higher-end GPU for better results.

Model Size

72.7B

Device VRAM

128 GB

Bandwidth

546 GB/s

Quantizations Tested

1

Performance by Quantization

Each row shows Qwen 2.5 72B Instruct performance at a different quality level on Apple M4 Max (128GB Unified).

QuantizationSpeedTTFTFits in VRAMRatingConfidence
Q4_K_M6 tok/s2200ms✓ YesAcceptableestimated

Notes

Q4_K_M

Q4 at 40.5GB in 128GB. Higher quality than Q3 with massive headroom.

About Qwen 2.5 72B Instruct

Qwen 2.5 72B Instruct (72.7B) is a chat, coding, ai coding, reasoning, multi-purpose model. Top-tier 72B model, competitive with Llama 3.1 70B. Excellent at coding and structured output. Requires 48GB+ VRAM for usable quantizations.

View all Qwen 2.5 72B Instruct hardware options →

About Apple M4 Max (128GB Unified)

Apple M4 Max (128GB Unified) has 128 GB at 546 GB/s. Available in MacBook Pro 16", Mac Studio.

See all models Apple M4 Max (128GB Unified) can run →

Source: MLX 70B+ model reports (2026-01-15)

Data last updated: 2026-03-01