GPU.ai homepage
Billed per second, from the moment an instance reaches SSH-ready until it is stopped, with no minimums and no idle reservation fees. On-demand rates vary by region and upstream provider, and the platform routes each launch to the cheapest one with stock. Reserved capacity is quote-only and arranged with sales.
Each card leads with a per GPU per hour rate. We prefer configurations the provider reports as in stock, on-demand rates before reserved or spot, then take the lowest per-GPU rate among those.
We match the GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.
GPU.ai lists 23 GPU models across 77 configurations, including the H100, B300 and B200. Prices start at $0.10 /GPU/hr on demand.
No GPU models match .
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Verify before provisioning. More on how we price. See GPU.ai's pricing.
GPU.ai is a compute aggregator. Rather than running its own datacenters, it scans live capacity and pricing across upstream GPU clouds and places each launch on the lowest cost host matching the requested spec.
Capacity spans North America, Europe, and India, and is exposed through a CLI, a REST API, and a web console. Researchers and ML teams can run on-demand GPUs without holding accounts at each upstream provider.
GPU.ai offers 1 of the 27 services we track.
| Service | Provider page |
|---|---|
| GPU-powered Servers | On GPU.ai |
Compare GPU.ai against other cloud providers.
Our data for GPU.ai was last updated on Oct. 1, 2026.
