Hardware for Local LLMs
All 21 GPUs and Apple Silicon devices tested for local AI inference. Use the VRAM calculator to see which models fit each device.
NVIDIA GPUs
NVIDIA GeForce RTX 4060 (8GB)
8 GB VRAM 115W
NVIDIA GeForce RTX 4070 (12GB)
12 GB VRAM 200W
NVIDIA GeForce RTX 4090 (24GB)
24 GB VRAM 450W
NVIDIA GeForce RTX 3090 (24GB)
24 GB VRAM 350W
NVIDIA GeForce RTX 5090 (32GB)
32 GB VRAM 575W
NVIDIA RTX 4080 (16GB)
16 GB VRAM 320W
NVIDIA RTX 4060 Ti (16GB)
16 GB VRAM 165W
NVIDIA RTX 5080 (16GB)
16 GB VRAM 360W
NVIDIA RTX 5070 (12GB)
12 GB VRAM 250W
NVIDIA RTX 5070 Ti (16GB)
16 GB VRAM 300W
NVIDIA RTX 4070 Ti Super (16GB)
16 GB VRAM 285W
NVIDIA DGX Spark
128 GB VRAM 170W
AMD GPUs
AMD Radeon RX 7900 XTX (24GB)
24 GB VRAM 355W
Intel GPUs
Intel Arc B580 (12GB)
12 GB VRAM 190W
Apple Silicon
Apple Mac mini (M4, 16GB)
16 GB VRAM 25W
Apple Mac mini (M4, 24GB)
24 GB VRAM 25W
Apple Mac mini (M4 Pro, 24GB)
24 GB VRAM 35W
Apple Mac mini (M4 Pro, 48GB)
48 GB VRAM 40W
Apple Mac Studio (M4 Max, 64GB)
64 GB VRAM 85W
Apple Mac Studio (M4 Max, 128GB)
128 GB VRAM 85W
Mini PC / eGPU
Beelink SEi12 + eGPU RTX 4090 (24GB)
24 GB VRAM 495W