4 bays · 2 x 2.5GbE
AINAS Score 23
No verified price

AI CapableOrico
Why AI Capable: Intel Celeron N5105/N5100 (no NPU), 8GB DDR4 in two SO-DIMM slots with no published maximum, no GPU path; Docker capable CyberData OS, so capable at best.
Updated
Can the Orico MetaBox Pro run local AI?
The Orico MetaBox Pro can run small local models, with limits. Its Intel Celeron N5105 and nullGB memory ceiling suit only very small models of around 1B to 3B parameters (Estimated), so treat it as a storage-first NAS that can host light AI tools.
There is no way to add a discrete GPU, so all inference runs on the CPU.
Local LLM support is documented by the manufacturer.
The MetaBox Pro suits a buyer who wants a five-bay home or small-office NAS with Docker and a way to try small AI apps, not a machine for serious local models. It scores 23 on the AINAS Score, which reflects that. It has an Intel Celeron N5105 with 4 cores, 8GB of DDR4 2666 memory in two SO-DIMM slots, and two 2.5GbE ports. Link aggregation is listed by the manufacturer. Two M.2 NVMe slots are present, which Orico describes as cache. CyberData OS runs Docker and has a virtual machine app. Orico's own guide confirms Open WebUI and Ollama as featured Docker apps, and recommends 1.5B models on lower end devices. The 120W power supply and 2-year warranty come from the manufacturer. Treat it as a storage box first, with AI as a small extra.
The main problem is AI capacity. Orico does not publish a maximum memory figure, there is no PCIe slot, OCuLink or USB4, and there is no NPU. Every model listed on this page is marked as not fitting at Q4 or Q8 (Estimated), starting with Llama 3.2 3B. That rules out buyers who want local LLMs, private document RAG or creative AI. It also affects anyone hoping to push Open WebUI beyond very small models. Those buyers should look for a system with a PCIe slot or OCuLink port for a GPU, or an integrated GPU or NPU with a stated TOPS rating. They should also look for a published memory maximum with upgradeable slots, and for more than 8GB of RAM included. The memory also has no ECC support, and networking stops at 2.5GbE.
No verified price
We have not verified a listing for the Orico MetaBox Pro at a retailer we track. The specifications below stand on their own; a price appears here as soon as a listing is confirmed.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon.co.uk or Amazon.com, as applicable, at the time of purchase will apply to the purchase of this product.
Each capability shows how it was confirmed: Verified by AINAS, Vendor confirmed, Community confirmed or Unverified.
| Capability | Supported | Evidence | Note |
|---|---|---|---|
| Local LLM | Limited | Vendor confirmed | ORICO CyberData OS guide deploys Open WebUI and pulls Ollama models; recommends 1.5B models on lower end devices. Store copy cites AI indexing and an AI-enhanced ecosystem. |
| Ollama | Yes | Vendor confirmed | Open WebUI featured app in CyberData OS Docker pulls models from Ollama.com. |
| Open WebUI | Yes | Vendor confirmed | Open WebUI is a Docker featured app in CyberData OS. |
| RAG | Limited | Vendor confirmed | ORICO guide notes Open WebUI includes a built-in RAG engine. |
| AI search | Limited | Vendor confirmed | CyberData OS Smart Album offers search by similar image. |
| GPU expansion | No | Vendor confirmed | No PCIe slot, OCuLink or USB4 listed. |
| Hardware acceleration | Limited | Unverified | Intel iGPU not listed by ORICO for this model; no NPU. |
| AI photos | Limited | Vendor confirmed | Smart Album with search by similar image documented for CyberData OS. |
| AI video | Limited | Unverified | No AI video feature documented for this model. |
| Surveillance AI | Limited | Unverified | No surveillance app documented. |
Calculated from the MetaBox Pro's memory and each model's published size at each quantization. Rows marked Estimated are calculations, not benchmarks. Check any model.
| Model | Best quantization | Runs? | Memory needed | Expected performance | Verification |
|---|---|---|---|---|---|
| Llama 3.2 3BLlama, 3.21B | Q3 | Marginal | 8GB | Not measured | Estimated |
| Llama 3.1 8BLlama, 8.03B | None fits | No | 14GB at Q4 | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | None fits | No | 15GB at Q4 | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | None fits | No | 18GB at Q4 | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | None fits | No | 20GB at Q4 | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | None fits | No | 19GB at Q4 | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | None fits | No | 25GB at Q4 | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | None fits | No | 29GB at Q4 | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | None fits | No | 31GB at Q4 | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | None fits | No | 60GB at Q4 | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | None fits | No | 89GB at Q4 | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | None fits | No | 65GB at Q4 | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | None fits | No | 96GB at Q4 | Not measured | Estimated |
One verdict per workload, from the MetaBox Pro's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
Poor for local LLM work. Local LLM support is limited and vendor confirmed, Ollama is vendor confirmed through Open WebUI, but every model listed on this page is marked as not fitting with 8GB of RAM and no GPU (Estimated).
Poor for private RAG. Open WebUI has a built-in RAG engine according to Orico, but RAG support is limited, and with the Celeron N5105 and no GPU path the vendor-confirmed pipeline has little hardware behind it.
AI search is poor here. The only documented feature is search by similar image in the CyberData OS Smart Album, which is vendor confirmed but limited, and the system has no NPU to help with wider indexing.
Poor for AI photos. Smart Album with search by similar image is documented and vendor confirmed as limited support, and the Celeron N5105 has no NPU or confirmed hardware acceleration, which is Unverified for this model.
We cannot confirm surveillance support, so the verdict is unknown. The capability row is marked limited but Unverified, and no surveillance app is documented for the MetaBox Pro, so treat it as unconfirmed.
Poor for a home lab. Docker and a virtual machine app are both present in CyberData OS, but the Celeron N5105 has 4 cores, memory is 8GB and there is no PCIe slot for expansion.
Poor for business use. The MetaBox Pro has no ECC memory and no 10GbE, though it offers five bays, a 2-year warranty and two 2.5GbE ports with link aggregation listed.
Poor for creative AI. GPU expansion is marked no and vendor confirmed, hardware acceleration is Unverified, and local LLM support is limited, so there is no route to dedicated graphics on the MetaBox Pro.
The MetaBox Pro supports up to an unstated amount ofGB of DDR4 2666 SO-DIMM (2 slots).
No. The MetaBox Pro has no PCIe slot or OCuLink port for a graphics card. Inference runs on the CPU, so keep to small models.
With limits. The MetaBox Pro can host a small RAG stack such as AnythingLLM or Open WebUI with a compact embedding model, but memory and compute restrict how large the answering model can be. It suits a personal document library more than a team knowledge base.
Networking is 2 x 2.5GbE. Fine for one or two users; indexing a large archive over the network will be slower.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 5-bay, 8GB | 8GB | None |
The MetaBox Pro scores best for Home Lab (28/100) and Business (28/100). It suits home lab owners who want one box for storage, containers, media and a few AI services.
It is a weaker fit for anyone who wants to run models larger than about 8B parameters and for buyers who expect to add a graphics card later.
Other AI Capable systems closest in overall score.
No. Llama 3.2 3B is the smallest model listed on this page, and it is marked as not fitting at Q4 or Q8 on the MetaBox Pro (Estimated). The same applies to every larger model listed. This is a calculation, not a benchmark. The system has 8GB of RAM included and no GPU path, so there is no way to add dedicated graphics memory. Orico's own guide recommends 1.5B models on lower end devices, which is a vendor recommendation rather than something we have tested.
Yes. The MetaBox Pro uses DDR4 2666 SO-DIMM memory in two slots, with 8GB included, rather than soldered memory. Orico does not publish a maximum capacity, so we cannot say how much it will accept. Treat any upgrade as untested until Orico confirms a limit. The system has no ECC support, so standard non-ECC modules are the only option to consider.
No. The MetaBox Pro has no PCIe slot, no OCuLink port and no USB4 or Thunderbolt ports, and the GPU expansion capability is marked no and vendor confirmed. The ports listed are USB 3.2 Gen1 Type-A x2 and Type-C x1. Orico does not list an Intel iGPU for this model, and hardware acceleration is Unverified. Buyers who need a GPU should look for a system with a PCIe slot or OCuLink.
Yes, CyberData OS, which is Orico's own system. It runs Docker and includes a virtual machine app, and its guide deploys Open WebUI as a featured Docker app that pulls models from Ollama.com, vendor confirmed. The same guide notes that Open WebUI includes a built-in RAG engine, though RAG support is limited. CyberData OS also lists link aggregation for the two 2.5GbE ports. Orico recommends 1.5B models on lower end devices, so treat AI apps on this 8GB system as small experiments rather than daily tools.
Yes. CyberData OS on the MetaBox Pro supports Docker, and Orico lists Open WebUI as a Docker featured app that pulls models from Ollama.com. Both are vendor confirmed. It also has a virtual machine app. The Celeron N5105 has 4 cores and the system has 8GB of RAM, so heavy containers or several virtual machines will be constrained, and local LLM support is marked as limited.
Yes. The MetaBox Pro has 2 M.2 M-Key NVMe 2280 slots. Orico describes them as cache, so do not assume they work as a general storage pool. They sit alongside 5 drive bays. They do not add memory for models or a GPU path, because NVMe storage is separate from the 8GB of RAM and the system has no PCIe slot.
Yes. Ollama support is documented by the manufacturer. See the capability table above for the source.
Yes. Open WebUI support is documented by the manufacturer. See the capability table above for the source.
Partly. AI photos support is documented by the manufacturer. See the capability table above for the source.
Partly. Surveillance AI support is not yet verified. See the capability table above for the source.
Partly. Hardware acceleration support is not yet verified. See the capability table above for the source.
It has 2 x 2.5GbE built in. The fastest port is 2 x 2.5GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.