GPU Leaderboard compares cloud GPU prices for AI inference
MarkTechPost has launched GPU Leaderboard, a free service that compares cloud GPU prices for AI inference, along with the GPU Leaderboard: Deploy It Chrome extension. The site collects on-demand and spot pricing from providers’ own pricing pages and ranks 201 live offers from 24 providers by cost per GPU-hour.
Its calculator estimates the price of a complete inference run using the selected accelerator, GPU count, duration and utilization. Users can also add a model to narrow the results. In the default example—8×H100 for 72 hours at 90% utilization—the run uses 576 GPU-hours. The cheapest listed price is $967.68 at Latitude, while the median across 15 providers is $1,716.48 and the highest is $3,545.28 at CoreWeave. The service also calculates effective cost per used GPU-hour, accounting for billed idle time.
Deploy It works on Hugging Face model pages, GitHub repositories and through a browser-toolbar search. It estimates VRAM requirements at FP16, INT8 and INT4, and displays suitable low-cost GPUs. The service includes provider information, regional availability and price-check history; daily refresh and coverage remain limitations compared with older trackers.
Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.