GPU VRAM Tracker for AI & LLMs

Compare cloud GPUs by total VRAM, price per GB of VRAM, and model-fit use case. Built for local LLM inference, Stable Diffusion, fine-tuning, and AI video workloads.

Last updated May 26, 2026 · Data refreshed every 6 hours
VRAM Tiers
7
Best $/GB
$0.001/GB-hr
Tracked Listings
5213

Start with the workload, then pick the VRAM tier

VRAM price tiers

VRAM Tier Best for Listings Best $/GB From
16GB VRAM Stable Diffusion, ~7B model inference, fine-tuning small models 720 $0.001/GB-hr $0.013/hr
24GB VRAM 13B model inference, SDXL, lightweight training 430 $0.001/GB-hr $0.021/hr
40GB VRAM 30B model inference, fine-tuning medium models 105 $0.002/GB-hr $0.080/hr
48GB VRAM 34B inference, multi-batch serving, image generation pipelines 217 $0.002/GB-hr $0.107/hr
80GB VRAM 70B model inference, fine-tuning large models 242 $0.005/GB-hr $0.400/hr
141GB VRAM 70B at full precision, multi-batch 70B serving 148 $0.004/GB-hr $0.600/hr
180GB VRAM 180B+ model inference, frontier model training 237 $0.004/GB-hr $0.686/hr
Answer pages