Cloudoora: GPU cloud instances now in our directory

We've added Cloudoora to LLM Hosting Pro. Cloudoora rents GPU cloud instances for running open-weight models yourself, and you can now compare its plans, prices and supported models alongside the other LLM hosts we track.

Who's behind Cloudoora

Cloudoora is a cloud hosting and infrastructure company headquartered in Helsinki, Finland. It was started by a developer who felt the hosting providers of the time didn't cater to developers starting out on their own projects. Most of its business is everyday hosting: shared and WordPress hosting, VPS and dedicated servers, domains and business email, all managed from one control panel with 24/7 support. It also offers one-click installs of open-source AI apps such as Open WebUI and Flowise on its VPS plans.

Cloudoora lists data centres in Finland, France, Germany, Sweden, the United Kingdom, the United Arab Emirates and the United States.

The hosting plans we've listed

Cloudoora offers 41 GPU instances on its GPU cloud page. Each one gives you a single GPU that you set up yourself, so you choose the model and the serving software. We've split them into two plans by GPU memory. Prices are from Cloudoora's site in US dollars and can change, so check there for current rates.

• Cloudoora GPU Cloud Instance - RTX & Workstation: NVIDIA consumer, workstation and L40-class GPUs with 6 to 48 GB of VRAM, from an RTX A2000 to an L40S. About US$105 to US$1,022 a month, or US$0.14 to US$1.31 an hour. It suits small and mid-size models, up to 70B-class models quantized on the 48 GB cards.
• Cloudoora GPU Cloud Instance - Data Centre: data-centre GPUs with 80 to 288 GB of VRAM, including NVIDIA A100, H100, H200, B200 and B300 and AMD MI300X. About US$511 to US$6,912 a month, or US$1.67 to US$9.47 an hour. It suits large models such as Llama 3.1 405B and Qwen3 235B-A22B.

Most of the GPUs can be billed by the hour with a one-hour minimum, or you can pay monthly or commit for 3 to 36 months. There's no setup fee.

Models

Because you bring your own model, Cloudoora can run any open-weight model that fits in the GPU's memory. In our directory we've linked the models that realistically fit each plan, across five model families: Gemma, Llama, Mistral, Qwen and gpt-oss. These range from Gemma 3 4B and Llama 3.1 8B on the smaller GPUs to gpt-oss-120b, Llama 4 Maverick and Llama 3.1 405B on the largest.

Compare Cloudoora with other providers on our LLM hosting plans page, or see the full Cloudoora host profile.