Loading tool details…
Honest pros, cons, and verdict on this ai infrastructure tool
✅ Per-minute billing with prices locked at launch
Starting Price
7 ₽/hour
Free Tier
No
Category
AI Infrastructure
Skill Level
Developer
gpu.nz is a Russian-language marketplace for renting NVIDIA GPU servers billed by the minute, with templates for ComfyUI, Ollama and vLLM plus an API, CLI and MCP server.
gpu.nz is a Russian-language service for renting NVIDIA GPU servers by the minute. It targets people running Stable Diffusion and ComfyUI, LLM inference, model training and rendering who need short-term access to hardware they do not own.
The catalog lists 102 GPU models across 67 countries, from consumer cards such as the RTX 3060, 3090, 4090 and 5090 to data-center GPUs including the A100, A800 and H100. Prices are shown before launch and locked for the life of the server. Billing is per minute at the hourly rate divided by 60; published starting rates include 7 rubles per hour for an RTX 3060, 24 for an RTX 4090, 136 for an A100 and 326 for an H100. Buyers can choose standard servers or cheaper interruptible servers that run at a bid rate and can be paused when outbid. Disks continue to be billed while they exist, and traffic is billed per gigabyte.
per minute
per minute
gpu.nz delivers on its promises as a ai infrastructure tool. While it has some limitations, the benefits outweigh the drawbacks for most users in its target market.
gpu.nz is a Russian-language marketplace for renting NVIDIA GPU servers billed by the minute, with templates for ComfyUI, Ollama and vLLM plus an API, CLI and MCP server.
Yes, gpu.nz is good for ai infrastructure work. Users particularly appreciate per-minute billing with prices locked at launch. However, keep in mind site, pricing and support are in russian and priced in rubles.
gpu.nz starts at 7 ₽/hour. Check their pricing page for the most current rates and features included in each plan.
gpu.nz is best for Running Stable Diffusion or ComfyUI workflows on rented GPUs and Short LLM inference or fine-tuning jobs with vLLM or Kohya SS. It's particularly useful for ai infrastructure professionals who need per-minute nvidia gpu rental.
There are several ai infrastructure tools available. Compare features, pricing, and user reviews to find the best option for your needs.
Last verified March 2026