LLM inference
vLLM, SGLang or Ollama with Open WebUI to serve and test language models.
GPU cloud for AI, LLMs and rendering
Rent NVIDIA GPU servers by the hour or by billing cycle. Compare live configurations and prices, start from a ready-made template and connect over SSH or the browser terminal.
- Live GPU inventory with current prices
- Templates for PyTorch, LLM serving and image generation
- SSH and terminal access in the browser
With ONIE GPU Cloud you rent a GPU server from the live inventory: choose the GPU model, VRAM and disk, start from a template such as PyTorch, vLLM, Ollama or ComfyUI, then connect over SSH. Hourly servers are charged from your account balance, and a stopped hourly server is only billed for its disk.
Available GPUs and current prices
Live inventory: availability and prices change with the market. The console confirms the final price before you order.
Loading current prices…
Templates ready for your workload
Fine-tuning
Unsloth Studio for fine-tuning language models on your own data.
Image and video generation
ComfyUI for Flux, SDXL and video workflows.
PyTorch and CUDA development
PyTorch and Ubuntu 24.04 images with NVIDIA CUDA, accessible over SSH.
Desktop or full VM
An Ubuntu desktop in the browser, or a full KVM virtual machine when you need one.
Each template lists its minimum CUDA version, VRAM and disk; only compatible GPUs are offered for it.
How to rent a GPU
- 01
Choose a template and disk size
- 02
Pick a compatible GPU from the live inventory
- 03
Confirm the price and create the server from your account balance
- 04
Connect over SSH or the browser terminal
How GPU billing works
- Hourly: the GPU is charged per hour while the server runs. A stopped hourly server is only charged for its disk.
- Billing cycles: monthly, quarterly, semi-annual or yearly, with any discount shown before you pay.
- Disk storage is charged per GB together with the GPU price.
- Prices are in USD; you can display them in VND at the current exchange rate.
Frequently asked questions
What can I run on a rented GPU?
Workloads that fit in the GPU memory: LLM inference and fine-tuning, image and video generation, rendering and CUDA development. Templates for PyTorch, vLLM, SGLang, Ollama, Unsloth and ComfyUI install the software for you.
How do I connect to the GPU server?
Over SSH with your key, or with the terminal in your browser from the console. Desktop and VM templates also provide graphical access.
Is the price fixed?
Prices come from the live GPU inventory and can change. The console shows the confirmed price before you order, and your server is billed at that confirmed price.
What do I pay when the server is stopped?
With hourly billing, a stopped server is only charged for its disk. Delete the server when you no longer need its data.
Which GPUs are available?
The table above lists the configurations currently available, including the NVIDIA RTX 5090 with 32 GB of VRAM when it is in stock. Availability changes throughout the day.