Cloud machines with a dedicated NVIDIA GPU, usually ready in a few minutes. JupyterLab in your browser, SSH, PyTorch and CUDA preinstalled. No contracts.
Create an account and deploy your first GPU machine today.
Get Started4
GPU Models
2
Tiers in Stock
$0.40/hr
Starting
NVIDIA GeForce RTX cards with 12 or 24 GB of VRAM, for notebooks, training, inference and rendering jobs.
RTX 3090
VRAM
24 GB
GPUs per machine
1
Max vCPUs
8
Max RAM
24 GB
In stock
Deploy →RTX 4090
VRAM
24 GB
GPUs per machine
1
Max vCPUs
8
Max RAM
32 GB
Temporarily unavailable
UnavailableNVIDIA RTX A6000 with 48 GB of VRAM, for larger models and bigger batches.
RTX A6000
VRAM
48 GB
GPUs per machine
1
Max vCPUs
12
Max RAM
64 GB
Out of stock
Unavailable