Vram.app - GGUF Memory & Speed Estimator

Model Configuration

Hugging Face GGUF URL (Optional)

OR SELECT PRESET

Pre-set Model / Fallback

GPU Hardware

Host Application (Overhead)

Context Length (Tokens) 8192

4k 32k 64k 128k

KV Cache Precision

100% GPU VRAM Fit

VRAM Usage 0.00 / 24.00 GB

50.0%

Weight VRAM 0.00 GB

KV Cache 0.00 GB

Host Overhead 0.00 GB

Total Required 0.00 GB

Estimated Decode Speed

0.0 tokens / sec