Per-hour pricing. Pay only for what you use.
Buy tokens in bulk and save. Use tokens to rent any GPU via API.
Common questions about GPU Hog pricing.
How does billing work?
The first hour is a prepaid minimum when you rent a GPU. After that, billing switches to per-minute at the GPU's hourly rate. If your balance hits zero, your pod is automatically paused (not terminated) and you get a 24-hour grace period to add more tokens. No surprise charges.
What are tokens?
Tokens are a prepaid balance you can use to rent GPUs via our API. Buy in bulk for up to 40% savings. Each token is worth 1 cent. Each GPU costs tokens equal to its hourly rate in cents - an RTX 4090 at $0.49/hr costs 49 tokens/hr, an H100 SXM at $3.49/hr costs 349 tokens/hr.
What happens if I run out of tokens?
If your token balance reaches zero, your running pods are automatically paused - not terminated - so your data stays safe. You then have a 24-hour grace period to add more tokens and resume your pod. Paused pods are billed at 5 tokens/hr for storage. If no tokens are added within 24 hours, the pod is terminated and data is deleted.
Can AI agents use GPU Hog?
Yes! GPU Hog supports HTTP 402 Machine Payable Protocol. AI agents can list GPUs, request a rental, pay via Stripe, and receive pod access - all programmatically. See our
Setup Guide for details.
What happens when my rental expires?
You will receive an email warning 30 minutes before expiry. When time is up, the pod is automatically terminated. Save your work to persistent storage before expiry.
Can I stop and restart a pod?
Yes. You can stop a running pod from your dashboard to pause billing, and restart it later. Your volume data is preserved while stopped.
What images are available?
Pods deploy with PyTorch 2.4, Python 3.11, CUDA 12.4 pre-installed. SSH, Jupyter, and HTTP access are enabled by default.