Skip to main content

​
GPU Instances

GPU Instances are full Linux containers with dedicated GPU hardware, root access, and a public IP address. They are ideal for training models, running experiments, fine-tuning, and any workload that needs direct access to GPU compute.

​
What You Get

Every GPU instance includes:
  • Root access — Full sudo privileges in an isolated Linux container.
  • Dedicated GPUs — One or more GPUs exclusively allocated to your instance.
  • Configurable specs — Choose your CPU cores, memory, and disk size.
  • Public IP address — Direct SSH access and the ability to expose services.
  • Pre-installed tools — CUDA drivers, Python, pip, and common ML libraries come ready to use.

​
Available GPUs

GPUVRAMBest ForStarting At
RTX 409024 GBInference, fine-tuning, small training runs~$0.35/hr
L40S48 GBLarge model inference, medium training~$0.80/hr
A10040 GBTraining, fine-tuning, research~$1.20/hr
A10080 GBLarge-scale training, multi-GPU workloads~$1.60/hr
H10080 GBMaximum performance training and inference~$2.50/hr
Prices shown are approximate starting rates and may vary based on availability and configuration. Check the Deploy page for current pricing.

​
Instance Lifecycle

Every instance moves through these states: 
configured → creating → deploying → deployed → stopping → stopped → terminated
StatusDescription
ConfiguredInstance configuration is saved but not yet submitted.
CreatingResources are being allocated.
DeployingThe container is starting up and installing software.
DeployedThe instance is running and accessible via SSH.
StoppingThe instance is shutting down.
StoppedThe instance is paused. Resources are still reserved.
TerminatedThe instance is permanently destroyed.

​
Billing

Instances are billed hourly based on the GPU type and configuration you select.
  • Billing starts when the instance reaches the deployed state.
  • Billing stops when the instance is terminated.
  • Charges are deducted from your credit balance automatically.
Terminating an instance permanently deletes all local disk data. There is no way to recover files after termination. Use Storage Volumes or transfer your files before terminating.

​
Next Steps

Deploy an Instance

Step-by-step deployment guide.

Connect via SSH

SSH, VS Code, and port forwarding.