OwnRig

Mixtral 8x7B Instruct on Apple M4 Max (36GB Unified)

Apple M4 Max (36GB Unified) handles Mixtral 8x7B Instruct well at 20 tok/s at Q4_K_M. A solid choice for this model.

Model Size

46.7B

Device VRAM

36 GB

Bandwidth

546 GB/s

Quantizations Tested

1

Performance by Quantization

Each row shows Mixtral 8x7B Instruct performance at a different quality level on Apple M4 Max (36GB Unified).

QuantizationSpeedTTFTFits in VRAMRatingConfidence
Q4_K_M20 tok/s500ms✓ YesGoodestimated

Notes

Q4_K_M

Q4 at 26.2GB fits in 36GB unified memory. MoE sparsity gives good speed despite total param count.

About Mixtral 8x7B Instruct

Mixtral 8x7B Instruct (46.7B) is a chat, coding, reasoning, multi-purpose model. Mixture of Experts model — 46.7B total params but only ~12.9B active per token. Excellent quality-to-speed ratio. Despite large total params, inference speed closer to a 13B model. Needs more VRAM for the full weight set though.

View all Mixtral 8x7B Instruct hardware options →

About Apple M4 Max (36GB Unified)

Apple M4 Max (36GB Unified) has 36 GB at 546 GB/s. Available in MacBook Pro 16", Mac Studio.

See all models Apple M4 Max (36GB Unified) can run →

Source: MLX MoE reports (2026-01-15)

Data last updated: 2026-03-01