Mac Studio M4 Max 36GB

Apple · M4 Max · 36GB Unified Memory · Can run 41 models

Buy Apple
Manufacturer Apple
Unified Mem 36 GB
Chip M4 Max
CPU Cores 14
GPU Cores 30
Neural Engine 16
Bandwidth 546 GB/s
MSRP $1,999
Released Mar 12, 2025

AI Notes

The Mac Studio M4 Max 36GB delivers incredible bandwidth at 546 GB/s, making smaller models fly. It runs 13B models at very fast speeds and can handle 30B models at moderate-to-fast speeds. The 36GB memory limits larger models, but the bandwidth advantage makes it excellent for smaller model inference.

Compatible Models

Model Parameters Best Quant VRAM Used Fit Est. Speed
Gemma 3 1B 1B Q8_0 2 GB Runs ~273 tok/s
Llama 3.2 1B 1B Q8_0 3 GB Runs ~182 tok/s
DeepSeek R1 1.5B 1.5B Q8_0 3 GB Runs ~182 tok/s
Gemma 2 2B 2B Q8_0 4 GB Runs ~137 tok/s
Llama 3.2 3B 3B Q8_0 5 GB Runs ~109 tok/s
Phi-3 Mini 3.8B 3.8B Q8_0 5.8 GB Runs ~94 tok/s
Phi-4 Mini 3.8B 3.8B Q4_K_M 4.5 GB Runs ~121 tok/s
Gemma 3 4B 4B Q4_K_M 5 GB Runs ~109 tok/s
Qwen 3 4B 4B Q4_K_M 4.5 GB Runs ~121 tok/s
DeepSeek R1 7B 7B Q8_0 9 GB Runs ~61 tok/s
Mistral 7B 7B Q8_0 9 GB Runs ~61 tok/s
Qwen 2.5 7B 7B Q8_0 9 GB Runs ~61 tok/s
Qwen 2.5 Coder 7B 7B Q8_0 9 GB Runs ~61 tok/s
DeepSeek R1 8B 8B Q4_K_M 7.5 GB Runs ~73 tok/s
Llama 3.1 8B 8B Q8_0 10 GB Runs ~55 tok/s
Qwen 3 8B 8B Q4_K_M 7.5 GB Runs ~73 tok/s
Gemma 2 9B 9B Q8_0 11 GB Runs ~50 tok/s
Gemma 3 12B 12B Q4_K_M 10.5 GB Runs ~52 tok/s
Mistral Nemo 12B 12B Q4_K_M 9.5 GB Runs ~57 tok/s
DeepSeek R1 14B 14B Q4_K_M 9.9 GB Runs ~55 tok/s
Phi-4 14B 14B Q4_K_M 9.9 GB Runs ~55 tok/s
Qwen 2.5 14B 14B Q4_K_M 9.9 GB Runs ~55 tok/s
Qwen 2.5 Coder 14B 14B Q4_K_M 12 GB Runs ~46 tok/s
Qwen 3 14B 14B Q4_K_M 12 GB Runs ~46 tok/s
StarCoder2 15B 15B Q8_0 17 GB Runs ~32 tok/s
Codestral 22B 22B Q4_K_M 14.7 GB Runs ~37 tok/s
Devstral 24B 24B Q4_K_M 17 GB Runs ~32 tok/s
Mistral Small 3.1 24B 24B Q4_K_M 18 GB Runs ~30 tok/s
Gemma 2 27B 27B Q4_K_M 17.7 GB Runs ~31 tok/s
Gemma 3 27B 27B Q4_K_M 20 GB Runs ~27 tok/s
Qwen 3 30B-A3B (MoE) 30B Q4_K_M 22 GB Runs ~25 tok/s
DeepSeek R1 32B 32B Q4_K_M 20.7 GB Runs ~26 tok/s
Qwen 2.5 32B 32B Q4_K_M 20.7 GB Runs ~26 tok/s
Qwen 2.5 Coder 32B 32B Q4_K_M 23 GB Runs ~24 tok/s
Qwen 3 32B 32B Q4_K_M 23 GB Runs ~24 tok/s
Command R 35B 35B Q4_K_M 22.5 GB Runs ~24 tok/s
Mixtral 8x7B 47B Q4_K_M 29.7 GB Runs ~18 tok/s
DeepSeek R1 70B 70B Q4_K_M 43.5 GB CPU Offload ~13 tok/s
Llama 3.1 70B 70B Q4_K_M 43.5 GB CPU Offload ~13 tok/s
Llama 3.3 70B 70B Q4_K_M 43.5 GB CPU Offload ~13 tok/s
Qwen 2.5 72B 72B Q4_K_M 44.7 GB CPU Offload ~12 tok/s
2 model(s) are too large for this hardware.