<Group> Help Create an account
VPS Quote Builder

Build your server and see the price instantly

Choose the processor, memory, disk and operating system. If you're going to run artificial intelligence, start from a use case: we'll suggest the right resources or GPU server.

Outbound traffic2 TB included
IP address1 public IPv4 included
ResourcesIsolated, just for you
PricesMonthly in USD, excl. VAT
Step 1 · Optional

What will you use your server for?

Pick a use case and we'll load a starting configuration. You can then adjust each resource.

Step 2

Configure the resources

Operating system

Linux is the most affordable option and the usual choice for AI, containers and web applications.

Processor

USD per month per vCPU
1163260

RAM

USD per GB per month
132128230

Disk

202501 TB4 TB

Network and extras

Additional outbound traffic2 TB included per month · from USD per additional TB, with lower per-TB rates at higher volumes. Inbound traffic is free.

0 TB

Reserved static IPv41 public IPv4 included at no cost, assigned automatically. If you need an address that never changes, reserve a static IPv4 for USD per month each.

0
View as Markdown
GPU Servers

Need a GPU? Choose a complete server

GPUs come in servers with processor, memory and disk already sized for artificial intelligence. The monthly price is fixed and includes the GPU.

Install the AI tools you already use

  • Ollama
  • ComfyUI
  • Open WebUI
  • vLLM
  • Hugging Face
  • PyTorch
  • LangChain
  • n8n
  • Dify
  • OpenClaw
  • Docker
  • Jupyter
  • NVIDIA CUDA
  • Fish Audio
  • Whisper
  • FFmpeg

Run state-of-the-art open models

  • Qwen3Text and code
  • Qwen-ImageImages
  • DeepSeekReasoning
  • GLM-5.2Agents and code
  • MiMoReasoning
  • Kimi K2Agents
  • LlamaText
  • Gemma 3Text and vision
  • MistralText
  • MiniMaxText and agents

Brands and logos belong to their respective owners and are shown only to indicate compatibility. Registro Digital is not affiliated with them. Each model has its own usage license.

Inferencia ligera

GPU T4

Modelos de 7 a 8 mil millones de parámetros, transcripción de audio y visión.

Ideal for
Ollama
Chat y agentes privados Qwen3 8B Llama 3.1 8B MiMo 7B
Whisper
Transcripción de audio Whisper large-v3
Fish Audio
Voz sintética y clonación de voz Fish Speech
GPU
1 × NVIDIA Tesla T4
Video memory
16 GB
Processor
AMD EPYC 7302P · 16 núcleos
RAM
128 GB DDR4 ECC
Storage
2 × 1 TB NVMe (RAID 1)
$997.10USD per month · excl. VAT

One-time setup: $997.10 USD excl. VAT

Request this server
El más elegido

GPU A10

Chatbots en producción con modelos de hasta 27 mil millones de parámetros comprimidos.

Ideal for
Ollama
Chat privado con Open WebUI Gemma 3 27B Qwen3 14B DeepSeek R1 14B
ComfyUI
Generación de imágenes Qwen-Image SDXL
vLLM
API compatible con OpenAI Qwen3 8B MiMo 7B
GPU
1 × NVIDIA A10
Video memory
24 GB
Processor
AMD EPYC 7313P · 16 núcleos
RAM
128 GB DDR4 ECC
Storage
2 × 960 GB NVMe
$1,267.50USD per month · excl. VAT

One-time setup: $1,267.50 USD excl. VAT

Request this server
Video e inferencia

GPU Flex 170

Transcodificación de video e inferencia con OpenVINO. No usa CUDA.

Ideal for
FFmpeg
Transcodificación y streaming de video
OpenVINO
Inferencia en GPU Intel Qwen3 8B Whisper
Visión artificial
Análisis de cámaras y video en lote
GPU
1 × Intel Data Center GPU Flex 170
Video memory
16 GB
Processor
Intel Xeon Gold 5412U · 24 núcleos
RAM
256 GB DDR5 ECC
Storage
2 × 1.92 TB NVMe
$1,335.10USD per month · excl. VAT

One-time setup: $1,335.10 USD excl. VAT

Request this server
Modelos grandes

GPU RTX PRO 6000

Modelos de 30 a 120 mil millones de parámetros, imágenes, video y ajuste fino.

Ideal for
Ollama / vLLM
Modelos grandes en un solo servidor Llama 3.3 70B Qwen3 32B gpt-oss-120b DeepSeek R1 70B
ComfyUI
Imágenes y video Qwen-Image FLUX Wan 2.2
Ajuste fino
LoRA sobre tus datos Qwen3 8B a 32B
GPU
1 × NVIDIA RTX PRO 6000 Blackwell
Video memory
96 GB
Processor
2 × Intel Xeon 6517P
RAM
256 GB
Storage
2 × 1.92 TB NVMe
$3,042.00USD per month · excl. VAT

One-time setup: $3,042.00 USD excl. VAT

Request this server
Máximo rendimiento

GPU H200 S

Modelos de 70 mil millones de parámetros con muchos usuarios a la vez.

Ideal for
vLLM
Alta concurrencia Llama 3.3 70B gpt-oss-120b Qwen3 32B
Ajuste fino
Modelos de hasta 30 mil millones Qwen3 32B
GPU
1 × NVIDIA H200
Video memory
141 GB
Processor
15 vCPU dedicadas
RAM
267 GB
Storage
1 TB
Location
Alemania (UE)
$3,094.69USD per month · excl. VAT

No setup fee

Request this server
2 GPU

GPU H200 M

Modelos de más de 100 mil millones de parámetros.

Ideal for
vLLM
Modelos de varios cientos de miles de millones Qwen3 235B MiMo-V2-Flash Llama 3.3 70B
Ajuste fino
LoRA sobre modelos de 70 mil millones Llama 3.3 70B
GPU
2 × NVIDIA H200
Video memory
282 GB (141 GB por GPU)
Processor
30 vCPU dedicadas
RAM
534 GB
Storage
1.5 TB
Location
Alemania (UE)
$6,188.43USD per month · excl. VAT

No setup fee

Request this server
4 GPU

GPU H200 L

Los modelos abiertos más grandes, comprimidos, y entrenamiento.

Ideal for
vLLM
Modelos de frontera comprimidos GLM-5.2 DeepSeek V3 Qwen3 235B
Entrenamiento
Entrenamiento distribuido en 4 GPU
GPU
4 × NVIDIA H200
Video memory
564 GB (141 GB por GPU)
Processor
60 vCPU dedicadas
RAM
1068 GB
Storage
2 TB
Location
Alemania (UE)
$12,377.81USD per month · excl. VAT

No setup fee

Request this server
8 GPU

GPU H200 XL

Los modelos abiertos más grandes con máxima calidad y entrenamiento.

Ideal for
vLLM
Modelos de frontera en 8 bits GLM-5.2 DeepSeek V3 Kimi K2
Entrenamiento
Modelos propios y varios modelos en paralelo
GPU
8 × NVIDIA H200
Video memory
1128 GB (141 GB por GPU)
Processor
127 vCPU dedicadas
RAM
2136 GB
Storage
4 TB
Location
Alemania (UE)
$24,755.61USD per month · excl. VAT

No setup fee

Request this server

GPU servers are activated subject to availability; we confirm the delivery date before payment. They include 2 TB of outbound traffic and 1 public IPv4.

AI guide

How much memory does your model need?

Approximate reference for open models such as Qwen, Llama, Mistral or DeepSeek. Actual usage depends on the model, quantization, context length and concurrent users.

Model sizeApproximate memory (4-bit quantized)Approximate memory (16-bit)Suggestion
Up to 3 billion parameters2 to 3 GB6 to 8 GBNo GPU · 4 vCPU · 8 GB RAM
7 to 8 billion5 to 6 GB16 to 18 GBNo GPU for testing · GPU T4 16 GB or GPU A10 24 GB in production
13 to 14 billion9 to 10 GB28 to 32 GBGPU A10 24 GB quantized · GPU RTX PRO 6000 96 GB at 16 bits
30 to 34 billion20 to 22 GB64 to 70 GBGPU RTX PRO 6000 96 GB quantized or at 16 bits
70 billion40 to 45 GB140 GB or moreGPU RTX PRO 6000 96 GB quantized · GPU H200 141 GB or more at 16 bits

Indicative values calculated from the number of parameters, plus a margin for context. We confirm the sizing before activating your server.

Frequently asked questions

Does the price include VAT?

No. Prices are shown excluding VAT; the tax is added at checkout.

What does each server include?

Each VPS is an isolated server with its own resources, 2 TB of outbound traffic per month, one public IPv4 at no cost and administrator access. Inbound traffic is free.

Does the included IPv4 change?

The included IPv4 is assigned automatically and may change if the server is fully shut down or rebuilt. If your domain, email or an integration needs an address that never changes, add a reserved static IPv4.

What is the difference between vCPUs and dedicated cores?

A vCPU shares the physical processor with other isolated servers and is ideal for most websites and applications. Dedicated cores reserve physical AMD EPYC cores just for you, with consistent performance for processes that keep the processor at full load all the time.

How is SQL Server billed?

SQL Server is licensed in 2-core packs, with a minimum of 2 packs per server. The Web edition may only be used for public websites, stores and APIs on the internet; for ERP, CRM or other internal applications, choose Standard or Enterprise.

How is the GPU billed?

GPUs are offered in complete servers with a fixed monthly price that already includes the GPU, processor, memory and disk. They cannot be added to a VPS built from individual resources.

Can I scale up resources later?

Yes. You can increase processor, memory, disk or traffic whenever your project needs it; the price adjusts to the new configuration.

Do I need a GPU to use artificial intelligence?

Not always. If your application uses external AI APIs or a small model for internal use, a processor and memory are enough. For medium or large models with fast responses and multiple users, a GPU server is the better choice.

Prefer to start from a plan?

Browse our VPS plans by category: general purpose, Windows, high memory, dedicated cores, artificial intelligence and storage.

View VPS plans What is a VPS?