Run Qwen 2.5 72B on M5 Ultra

Yes — Qwen 2.5 72B (72B) runs at 23 tok/s on M5 Ultra with 96 GB RAM using Q4_K_M quantization via MLX. First token latency is Nones. Alibaba's 72B Qwen 2.5 model — powerful but requires 64 GB+ Mac.

Speed
estimated23
tok/s
First Token
estimatedNone
seconds
RAM Needed
96
GB minimum
Engine
MLX
recommended
Want to run Qwen 2.5 72B faster — or step up to bigger models? Find your Mac →

Benchmark Details

The LLMCheck index estimates Qwen 2.5 72B on M5 Ultra using our published methodology: Q4_K_M quantization, memory-bandwidth scaling, and cross-referenced third-party benchmarks where available. Figures are transparent estimates — own this config? Submit a real benchmark →

MetricValue
Tokens per second23 tok/s
Time to first tokenNones
QuantizationQ4_K_M
Minimum RAM96 GB
Recommended engineMLX
Parameters72B
Benchmark date2026-08

Q4_K_M 72B MLX M5 Ultra

Setup Guide: Run Qwen 2.5 72B on M5 Ultra

The recommended engine for Qwen 2.5 72B on M5 Ultra is MLX. Install with pip and pull the model:

pip install mlx-lm
mlx_lm.generate --model mlx-community/qwen-25-72b-q4_k_m --prompt "Hello!"

Alternatively, you can use Ollama for a simpler setup:

ollama run qwen2.5:72b

Performance on Other Apple Silicon Chips

ChipSpeedFirst TokenMin RAMEngine
M4 Ultra 21 tok/s 2.5s 192 GB Ollama
M5 Max 12 tok/s 3.2s 128 GB Ollama

System Requirements

To run Qwen 2.5 72B on M5 Ultra you need:

🛒 Get a Mac that runs Qwen 2.5 72B

Qwen 2.5 72B needs about 96 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:

As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep our benchmarks free.

Compare More Models

See how Qwen 2.5 72B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard