from 24 GBChat
Private ChatGPT · Qwen3 32B
A chat in your browser for writing, code and documents. No subscriptions and no message limits.
from €0.30/h · RTX 3090, 24 GB
Run AI models · DeepSeek
Run DeepSeek R1 on a rented GPU
The 7B and 14B DeepSeek R1 distills run on a 12 GB card, and the 32B version needs about 23 GB in 4-bit. The full 671B R1 needs hundreds of gigabytes of VRAM, so in practice people run the distills: they fit on one or two cards.
- App
- PyTorch
- GPU
- RTX 3060, 12 GB
- Price
- from €0.12/h
To launch, your balance must cover one hour of the machine. You are billed by the minute for actual use.
from 5 GBof video memory
2–5 minutesuntil the server is up
By the minutebilled from your balance
SSH and Jupyterserver access
How much memory you need and which GPU to pick
R1 Distill 7B takes about 5 GB of video memory in 4-bit; the cheapest suitable card is RTX 3060. Prices update from available catalogue machines when you open the page.
| Version | FP16 | 4-bit | Suitable GPUGPU | Price per hourPer hour |
|---|---|---|---|---|
| R1 Distill 7B | 16 GB | 5 GB | RTX 3060 | €0.12 |
| R1 Distill 14B | 31 GB | 10 GB | RTX 3060 | €0.12 |
| R1 Distill 32B | 71 GB | 22 GB | 2× RTX 3060 | €0.24 |
| R1 Distill 70B | 154 GB | 47 GB | 4× RTX 4060 Ti | €0.64 |
| R1 / V3 full (671B, MoE) | 1477 GB | 443 GB | 8× A100 | €11.68 |
Estimate: FP16 ≈ 2 GB per billion parameters, 4-bit ≈ 0.6 GB, plus 10% for context. Pick: the cheapest configuration with enough memory; for large models that means several identical cards. All GPUs and prices
How to run DeepSeek in 3 steps
- Sign up and top up. Create an account and top up from €10. To launch, your balance must cover at least one hour of the chosen machine; after that you pay by the minute.
- Choose a GPU. Memory for each version is in the table above, and the lightest one runs from 5 GB. The button on this page selects the cheapest option with the PyTorch template.
- Run the model. The server is ready in 2–5 minutes. Open Jupyter in your browser or connect over SSH and run the commands below.
Finished? Stop the server in the dashboard and billing stops. Save and stop keeps /workspace for €0.03 per GB per month, and it comes back on your next server automatically.
With Ollama (easiest)
curl -fsSL https://ollama.com/install.sh | sh
ollama run deepseek-r1:32b
With vLLM (API, heavy load)
pip install vllm
vllm serve deepseek-ai/DeepSeek-R1-Distill-Qwen-32B --max-model-len 16384
What people run DeepSeek for
01
Reasoning tasks
Maths, logic, finding a bug in code. R1 writes out its chain of thought before the answer, so you can check it.
02
Search over your documents
RAG on top of contracts, manuals or a knowledge base: the model answers and points to the right passage.
03
An API for a bot or a website
vLLM exposes an OpenAI-compatible API, so you only need to point your client code at your own server.
Tips before you launch
- The 32B distill in 4-bit is the best balance of quality and price: it fits on a single RTX 3090 or 4090.
- R1 reasons at length, so plan for a 16–32k token context. That adds a few gigabytes of memory.
- The full 671B only makes sense in 4-bit on 8× H100 or H200, which costs tens of euros an hour.
Just need a chat? A ready-made app with no setup
If you need the model for conversations and don’t want to write code, launch Private ChatGPT: the server comes with Ollama, a Qwen3 model and the Open WebUI chat window in your browser.
from 12 GBChat
Private ChatGPT · Qwen3 8B
A fast chat on an affordable card: email drafts, translations, quick answers.
from €0.12/h · RTX 3060, 12 GB
Frequently asked questions
Which GPU do I need for DeepSeek R1?
An 8–12 GB card is enough for the 7B distill, 24 GB (RTX 3090 or 4090) for 32B in 4-bit, and 48 GB (RTX A6000, L40S) or two 24 GB cards for 70B in 4-bit.
Can I run the full DeepSeek R1 671B?
Yes, but it needs about 400 GB of VRAM even in 4-bit. That means 8× H100 or H200.
How is a distill different from the full model?
Distills are Qwen and Llama models fine-tuned on R1’s reasoning. They are weaker than the original but many times cheaper to run.
How long does a launch take?
The server is usually ready in 2–5 minutes, then SSH and Jupyter are available. Downloading the model depends on its size and is billed as normal server time.
What happens to my models and results after I stop?
When you stop, click Save and stop: the /workspace folder with your models, LoRAs and results goes to storage for €0.03 per GB per month and comes back on your next server automatically. If you stop without saving, the disk is deleted.
How is it billed?
By the minute from your balance in euros. Top up from €10; to launch, your balance must cover at least one hour of the chosen machine. Stop the server and billing stops.
A server is usually ready in 2–5 minutes
Sign up in a minute, top up from €10, pay by the minute. Stop the server and billing stops.