Question 1

Which plan for a first local model?

Accepted Answer

Core VPS 4 at €12 with Ollama covers 7–8B quantized models comfortably. If you want 13–14B or several models side by side, take Core VPS 6 with 16 GB.

Question 2

Is there an OpenAI-compatible API?

Accepted Answer

Yes — llama.cpp server and vLLM both expose OpenAI-style endpoints. Point your app's base URL at the VPS and keep working with the code you already have.

Question 3

Do you provide GPUs?

Accepted Answer

No. All Buran VPS lines are CPU + NVMe. For training and GPU inference you need a specialized provider — we say so before you order, not after.

🚀 VPS for AI — Run LLMs (Ollama, llama.cpp) on Your Own Server