Marktechpost has launched GPU Leaderboard, a free cloud GPU price comparison for AI inference. It also shipped GPU Leaderboard: Deploy It, a Chrome extension for the same data. The site reads on-demand and spot prices from each provider’s own pricing page. The goal is simple: answer “what GPU do I need, and where is it cheapest?” in one place.
What is GPU Leaderboard?
GPU Leaderboard is a pricing dashboard that ranks rentable cloud GPUs by price per GPU-hour. It is built for teams running inference.
The v1 application has 4 views:
- Inference pricing: a cost calculator for a full run, not a single hourly rate.
- Leaderboard: every offer ranked by $/GPU-hr, filterable by GPU, tier, region and availability.
- Cheapest Inference: the lowest on-demand floor for each tracked GPU.
- Providers: a directory with instance counts, regions, status-page incidents and price-check history.
How does the GPU inference cost calculator work?
You pick an accelerator, GPU count, duration and utilization. The calculator then prices that exact run at every provider. An optional model field narrows the list.
The default example is 8×H100 for 72 hours at 90% utilization. That bills 576 GPU-hours. The cheapest result is $967.68 at Latitude, at $1.68 per GPU-hour. The median across 15 providers is $1,716.48. The most expensive is $3,545.28 at CoreWeave.
The calculator also shows a “real” cost per used GPU-hour. Idle time is still billed, so $1.68 at 90% utilization becomes $1.87. This is the number most budget sheets miss.