Train and run any model. On a real GPU
From 16GB to 80GB VRAM available. ComfyUI, Ollama, vLLM, PyTorch — root access, CUDA pre-installed, no shared notebooks.
Full admin access means you can install anything. Here are some common stacks you can run.
ComfyUI
Node-based Stable Diffusion workflows with full GPU acceleration.
Automatic1111
The Stable Diffusion web UI. Install it and run it on your GPU.
vLLM
Serve open-weight LLMs behind an OpenAI-compatible API.
Ollama
CLI LLM runner. Pull and run models from the command line.
Fooocus
Simplified Stable Diffusion — great for fast iteration.
KoboldCpp
Run GGUF models with GPU offloading for text generation.
InvokeAI
Professional Stable Diffusion toolkit with a polished UI.
LoRA training
Fine-tune diffusion models on your own data.
Choose the right tier based on what you want to run.
| model | vram needed | starter 16GB | standard 24GB | pro 48GB | power 80GB |
|---|---|---|---|---|---|
| Stable Diffusion XL | 8–12 GB | yes | yes | yes | yes |
| Flux.1 (dev/schnell) | 12–24 GB | tight | yes | yes | yes |
| Llama 3 8B (Q4) | ~6 GB | yes | yes | yes | yes |
| Llama 3 70B (Q4) | ~40 GB | no | no | tight | yes |
| Mixtral 8x7B (Q4) | ~26 GB | no | tight | yes | yes |
| SDXL LoRA training | 16–24 GB | tight | yes | yes | yes |
Standard
- gpu
- RTX 4090
- vram
- 24 GB
great for SD and smaller LLMs
Pro
- gpu
- L40S
- vram
- 48 GB
more VRAM for bigger models
Power
- gpu
- A100 80GB
- vram
- 80 GB
run 70B+ models and big batches
Ready to get started?
Create an account in seconds and deploy your first GPU box today.