Skip to main content

GPU Pricing

TensorPool charges per GPU/hour with no separate charges for vCPUs or memory. You can view pricing for all instance types at tensorpool.dev/pricing. All charges are prorated to the second. You only pay for the amount of time your resources are active. Pricing is subject to availability. In the event TensorPool’s primary capacity pool for an instance type is fully utilized, a fallback capacity pool may be available at a higher cost. If this occurs, you will be informed when you are creating the cluster and you always have the option to destroy the cluster.

Storage Volumes

TypePrice
Fast Volume$0.10/GB/month ($100/TB/month)
Flex Volume$0.07/GB/month (peak usage) + $0.005 per 1,000 requests
Flex storage uses request-based pricing because it is backed by object storage. Requests include standard S3 operations: GET, PUT, POST, COPY, LIST, SELECT, and DELETE. Since Flex storage volumes do not have a capped volume size, they are billed based on the peak size of the volume during the month. For example, if you store 100GB at any point during the month, you are billed for 100GB even if you delete the data later.

Billing

View your billing and usage at https://tensorpool.dev/dashboard/balance.