LeaderGPU is still beta in this directory. Headquarters on file is the Netherlands. We do not have a founding year. Billing is per minute, not per hour. Payment methods on file are credit card, PayPal, and crypto. The published catalogue includes H200 SXM, H100 SXM, A100, L40S, RTX A6000, A40, RTX 5090, RTX 4090, RTX 3090, and RTX 6000 Ada.
Why LeaderGPU stands out
The beta label is real, and the feature card is thin. Our data records persistent storage, multi-GPU, and Docker. An API, Jupyter, Kubernetes, spot, reserved instances, and SOC 2 are not set. What is no longer true is “no inventory.” Twelve GPU rows are published. Read them as per-minute rates.
Pros
- Published datacenter and GeForce rows, including H200, H100, 4090, and 5090
- Per-minute billing, with credit card, PayPal, and crypto
- Docker and persistent storage recorded
- Netherlands headquarters on file
Cons
- Beta. This is not a fully promoted listing
- No API, Jupyter, or Kubernetes in our data
- No spot, no reserved instances, and no SOC 2 flag
- No founding year
Getting started
- Open LeaderGPU and confirm the price is per minute before you multiply it out.
- Pay by credit card, PayPal, or crypto.
- Plan on Docker. An API and Jupyter are not recorded.
Best for: A per-minute dedicated GPU, including H100 or a 4090, if you can work without an API or Jupyter and you accept the beta label.