Skip to main content

Decision Tools

VRAM Calculator

Find out how much GPU memory you need to run any open-weight model, and which models fit your hardware.

Runs locally

Model size

7B
1B999B

= 7B parameters

Quantization

GPU budget

24 GB
4 GB192 GB

VRAM required

4 GB

7B @ INT4 / Q4

GPU budget

24 GB

available

Headroom

+20 GB

fits comfortably

✓

Fits on your GPU with 20 GB to spare

7B params × 0.5 bytes × 1.18 overhead = 4 GB

Open-weight models that fit (0)

No open-weight models fit within 24 GB at INT4 / Q4.