Run Qwen 4 Preview 32B-A3B on M5 Max

Yes — Qwen 4 Preview 32B-A3B (32B) runs at 78 tok/s on M5 Max with 128 GB RAM using Q4_K_M quantization via MLX. First token latency is 0.5s. A capable open-source LLM with 32B parameters.

Speed
78
tok/s
First Token
0.5
seconds
RAM Needed
128
GB minimum
Engine
MLX
recommended

Benchmark Details

LLMCheck measured Qwen 4 Preview 32B-A3B on M5 Max using the standard methodology: Q4_K_M quantization, 256-token input, 512-token output, 3 runs averaged on a freshly-booted system.

MetricValue
Tokens per second78 tok/s
Time to first token0.5s
QuantizationQ4_K_M
Minimum RAM128 GB
Recommended engineMLX
Parameters32B
Benchmark date2026-05

Q4_K_M 32B MLX M5 Max

Setup Guide: Run Qwen 4 Preview 32B-A3B on M5 Max

The recommended engine for Qwen 4 Preview 32B-A3B on M5 Max is MLX. Install with pip and pull the model:

pip install mlx-lm
mlx_lm.generate --model mlx-community/qwen-4-preview-32b-a3b-q4_k_m --prompt "Hello!"

Alternatively, you can use Ollama for a simpler setup:

ollama run qwen-4-preview-32b-a3b

Performance on Other Apple Silicon Chips

ChipSpeedFirst TokenMin RAMEngine
M4 Max 60 tok/s 0.7s 128 GB Ollama
M4 Pro 58 tok/s 0.8s 24 GB Ollama
M5 Pro 52 tok/s 0.7s 64 GB MLX
M3 Max 45 tok/s 0.9s 64 GB Ollama
M3 Pro 32 tok/s 1.2s 18 GB Ollama

System Requirements

To run Qwen 4 Preview 32B-A3B on M5 Max you need:

Compare More Models

See how Qwen 4 Preview 32B-A3B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard