Cloudoora GPU Cloud Instance - Data Centre
Rent Cloudoora data-centre GPUs with 80 to 288 GB of VRAM, from A100 to B300 and AMD MI300X, and run large open models yourself.
Overview
Cloudoora's large-memory GPU instances give you a single data-centre GPU with 80 to 288 GB of VRAM, for serving or fine-tuning large open-weight models with your own software stack. Options include NVIDIA A100, H100, H200, B200, B300 and RTX PRO 6000, plus AMD MI300X. Prices run from about US$511 (MI300X) to US$6,912 a month, and most of these GPUs can also be billed by the hour (US$1.67 to US$9.47) with a one-hour minimum. Longer terms of 3 to 36 months are available, with no setup fee.
Specifications
| Hosting type | GPU cloud |
|---|---|
| Model families | Llama, Qwen, gpt-oss |
| LLM models | Llama 3.3 70B Instruct, Qwen2.5 72B Instruct, gpt-oss-120b, Qwen3 235B-A22B |
| Pricing range | US$1.67 – US$9.47 |
| Pricing model | Per GPU-hour, Monthly subscription |
| Free tier available | No |
| Data center locations | Finland, France, Germany, Sweden, United Arab Emirates, United Kingdom, United States |
| Hardware / GPU options | Single-GPU instances: A100 SXM and H100 PCIe (80 GB), RTX PRO 6000 and RTX PRO 6000 WK (96 GB), H200 NVL (143 GB), B200 (180 GB), AMD MI300X (192 GB) and B300 (288 GB). |
| VRAM & memory | 80 GB – 288 GB |
| Maximum model size | 405B |
| OpenAI-compatible API | No |
News about Cloudoora
All news →Cloudoora: GPU cloud instances now in our directory
We've added Cloudoora, the Helsinki-based hosting company, with two GPU cloud plans for running open models yourself, from an RTX A2000 up to an NVIDIA B300.
