I Ran DeepSeek V4.1 at MAX on a Local AI Cluster (It’s Terrifying)

DeepSeek are back with more insane benchmarks, topping the charts in coding even beating Opus 5. So let's run it locally at MAX to see how good it really is. TEST SYSTEM Inferencer v2.4.4: https://inferencer.com Q4i: https://huggingface.co/inferencerlabs/DeepSeek-V4.1-MLX-Q4i Q9: https://huggingface.co/inferencerlabs/DeepSeek-V4.1-MLX-Q9 M3 Ultra 512GB + M4 Max 128GB: https://vtudio.com/a/?a=mac+studio BUY NOW Mac Studio: https://vtudio.com/a/?a=mac+studio MacBook Pro: https://vtudio.com/a/?a=macbook+pro LG C2 42" Monitor: https://vtudio.com/a/?a=lg+c2+42 HDD Drive: https://vtudio.com/a/?a=wd+