Comment by frwickst

I'm getting 6.55t/s using the Qwen3.5-397B-A17B-4bit model with the command: ./infer --prompt "Explain quantum computing" --tokens 100

MacBook Pro M5 Pro (64GB RAM)