Olares One: local LLM server
Last updated
Olares One is a server box listed at 4,999 USD with an RTX 5090 Mobile (24 GB VRAM), 96 GB of RAM and Olares OS. We feature it as a local LLM server for models up to about 32B shared by several people.
Olares One → Local LLM server → vLLM · Ollama · OpenAI-compatible API → Your models, served on your network
What does Olares publish about the box?
Olares publishes Olares One as a 4,999 USD box built around an NVIDIA RTX 5090 Mobile with 24 GB of GDDR7 VRAM. The rest of the server is an Intel Core Ultra 9 275HX, 96 GB of DDR5-5600 RAM, a 2 TB NVMe SSD and Olares OS, an open-source system built on Kubernetes.
- RAM upgradable to 128 GB; a second M.2 slot free
- 2.5 GbE, Thunderbolt 5 (80 Gbps), Wi-Fi 7
- Olares Market lists Ollama and Open WebUI as one-click installs
- Multi-user accounts, a built-in VPN, encrypted periodic backups to external locations
- 330 W power adapter, 30 W standby, 320 x 197 x 55 mm, 2.15 kg
Why is it featured as a local LLM server?
Olares One is featured because Olares published speed per request under load, the figure a server buyer needs most. In Olares' own test with vLLM, Qwen3-30B-A3B ran at about 81 tokens per second per request with 8 concurrent requests. With Ollama the same box gave 32 tokens per second per request at 8, which shows how much the engine matters.
The rest of the server stack comes with the OS. A small team gets accounts, remote access and backups without assembling a reverse proxy, a VPN and a login system on bare Linux.
How do you set it up?
Setting up Olares One as a model server takes one online activation and a few installs. Activation needs an internet connection, an Olares ID and the LarePass app, according to Olares' FAQ. Then you install Ollama and Open WebUI from the Olares Market, load a model that fits in 24 GB and create an account for each user.
Users connect through LarePass, locally or remotely. Apps connect through the engine's OpenAI-compatible endpoint, described in our API guide.
What should you confirm with Olares directly?
Confirm with Olares how to run vLLM on the box for multi-user serving, since the Market listing we checked names Ollama and Open WebUI. Ask what the box sends to Olares during activation, updates and remote access. The disk is not encrypted by default, according to Olares' FAQ, so plan for that. Then confirm the price at checkout.
The shop announced a 300 USD discount per unit starting 17 Nov 2026. Shipping is free to the US, the EU, the UK, Canada, Japan, South Korea and Hong Kong, per Olares' shipping policy.
Questions & answers
Can Olares One run without Olares OS?
Olares says you can dual-boot Windows, and Windows apps also run in a VM from the Olares Market. The multi-user and remote access features described here come from Olares OS.
Can I connect two Olares One units?
Yes. Two or more Olares devices can form a cluster, set up from the command line today. It raises throughput and allows larger models through vLLM pipeline parallelism, but one request doesn't get faster.
Sources for this entry
- Olares One product page (Company-published)
- Olares One specifications (Olares docs) (Company-published)
- Olares One FAQ (Olares docs) (Company-published)
- Olares blog: local AI hardware benchmarking (Oct 2025) (Company-published)
- Olares on GitHub (read via GitHub API 13 Nov 2026) (Company-published)
Continue this reading path
How the related websites connect
Cite this entry
Local LLM Server. “Olares One: local LLM server.” Content reviewed 13 November 2026.
https://localllmserver.com/olares-one/