GPU Per Hour

Real-time aggregator comparing cloud GPU prices across 25+ providers for AI and ML workloads.

GPU Per Hour is a real-time cloud GPU pricing aggregator and comparison platform designed for AI researchers, developers, and data scientists.

It tracks over 10,000 GPUs across 25+ cloud providers, including AWS, Lambda Labs, RunPod, and Vast.ai. The platform allows users to compare hourly rates for high-performance hardware like the NVIDIA H100, A100, and RTX 4090. It features advanced filtering options for VRAM, NVLink, security tiers, and deployment types, alongside a cost calculator and head-to-head provider comparisons to help users find the most cost-effective compute resources for machine learning and rendering tasks.

Pros

Real-time pricing and availability tracking for 10,000+ GPUs
Head-to-head comparisons between 25+ cloud providers
Advanced filtering by GPU model, VRAM, NVLink, and security tier
Cost calculator to estimate true hourly expenses
Coverage of both enterprise (H100/A100) and consumer (RTX 4090/3090) GPUs

Cons

Finding the lowest hourly rate for an NVIDIA H100 SXM instance
Comparing the cost and features of AWS versus specialized providers like RunPod
Locating available RTX 4090 stock for distributed machine learning training

DOWNLOAD

GPU Per Hour

unknown