4 x M.2 · 2 x 2.5GbE
AINAS Score 20
£281.31
Amazon UK · Verified

AI CapableTerraMaster
Why AI Capable: Intel Celeron N5095 (4 cores, UHD Graphics, no NPU) with 4GB shipped and 16GB max DDR4 and no GPU path: Docker-capable x86 but only tiny models.
Updated
Can the TerraMaster F2-425 run local AI?
Partly. The TerraMaster F2-425 runs local models on Intel Celeron N5095 alone, and its 16GB ceiling allows 7B to 8B class models (Estimated), but replies will be slow without graphics acceleration.
No graphics card can be fitted, which keeps every model on the CPU.
Local LLM support is not yet verified.
The F2-425 suits a buyer who wants a small two-bay NAS with a 2.5GbE port, Docker and a few light AI-flavoured apps, not a local LLM machine. It has an Intel Celeron N5095 with 4 cores, 4GB of DDR4 memory fitted and room for up to 16GB. TOS 7 runs on it, with Docker and virtual machines supported. The Photos app offers AI face and scene recognition (Vendor confirmed), and Surveillance Manager handles up to 32 ONVIF IP cameras (Vendor confirmed). Intel UHD Graphics handles video transcoding up to 4K60. The AINAS Score is 31, which reflects a system aimed at storage first. It carries a 2 year warranty and runs from a 48W power supply.
The main thing that rules it out is the lack of any route to more AI capability. There is no PCIe slot, no USB4, no Thunderbolt and no OCuLink, so a graphics card cannot be added. Memory tops out at 16GB, and there is no NPU. That affects anyone who wants to run models of 12B parameters or more, or who wants usable local chat and document search. Qwen3 8B is the largest model on this page that fits at Q4 (Estimated), and the 12B class is marginal at best. Buyers in that group should look for a system with at least 32GB of memory, a PCIe slot or OCuLink for a GPU, or an NPU.
Amazon UK
4GB
£279.99
Verified
In stock
Buy on AmazonProduct 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 | Unverified | Low-power CPU with small shipped RAM and no GPU path; only very small models via Docker would be practical. TerraMaster blog describes Ollama/Open WebUI via Docker on x86 TOS NAS generally, not this model. |
| Ollama | Limited | Vendor confirmed | Runs as a Docker container via TOS Docker Manager per TerraMaster blog (x86 TOS NAS generally). Practical only for tiny models on this hardware. |
| Open WebUI | Yes | Vendor confirmed | Open WebUI container via Docker; same TerraMaster blog source. |
| RAG | Limited | Unverified | Possible through Open WebUI or similar containers; no TerraMaster or community source showing RAG on TOS. |
| AI search | Limited | Vendor confirmed | TOS 7 OpenClaw agent and AI tagging allow natural-language queries; setup uses an API key from a provider such as OpenAI, so cloud LLM. Global search is an inverted index, not semantic. |
| GPU expansion | No | Vendor confirmed | Spec lists no PCIe slot; no USB4, Thunderbolt or OCuLink. |
| Hardware acceleration | Yes | Vendor confirmed | Hardware transcoding engine H.264/H.265/MPEG-4/VC-1 up to 4K60 via Intel iGPU. Video only, no NPU. |
| AI photos | Yes | Vendor confirmed | TOS 7 Photos: AI face recognition and AI scene recognition (80+ categories). |
| AI video | No | Unverified | No AI video feature found in TOS 7 material. |
| Surveillance AI | Yes | Vendor confirmed | TOS 7 Surveillance Manager supports up to 32 ONVIF IP cameras. |
Calculated from the F2-425'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 | Q8 | Yes | 13GB | Not measured | Estimated |
| Llama 3.1 8BLlama, 8.03B | Q5 | Yes | 15GB | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | Q4 | Yes | 15GB | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | Q4 | Marginal | 15GB | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | Q4 | Marginal | 16GB | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | Q4 | Marginal | 16GB | 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 F2-425's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The F2-425 is a poor choice for running a local LLM. Local LLM support is limited and Unverified, and Ollama is limited too, so only very small models via Docker are practical on this CPU with no GPU path.
Private RAG is a poor fit on the F2-425. RAG support is limited and Unverified, with no TerraMaster or community source showing it on TOS, though Open WebUI is Vendor confirmed as a Docker container.
AI search is a poor fit here. The TOS 7 agent and AI tagging need an API key from a cloud provider, and the built-in global search is an inverted index, not semantic, so nothing local is doing the AI work.
AI photos is rated poor, even though the feature exists. TOS 7 Photos offers AI face recognition and scene recognition across 80+ categories (Vendor confirmed), but the Celeron N5095 has no NPU to speed it up.
Surveillance is rated poor on the F2-425. Surveillance Manager supports up to 32 ONVIF IP cameras (Vendor confirmed), but there is no NPU and no AI video feature in the TOS 7 material.
Home lab use is rated poor. Docker and virtual machines are both supported, but 4GB of memory fitted, a 16GB ceiling, a four-core Celeron and no M.2 slots leave little room for several services.
Business use is rated poor. The F2-425 has two bays, non-ECC memory and a single 2.5GbE port, which gives little headroom for shared workloads, though it does carry a 2 year warranty.
Creative AI is a poor fit. GPU expansion is not supported (Vendor confirmed), and hardware acceleration covers video transcoding only through the Intel iGPU, with no NPU and limited, Unverified local LLM support.
Out of the box the F2-425 has 4GB; the official ceiling is 16GB of DDR4 non-ECC SODIMM. A memory upgrade is the cheapest way to make it more useful for AI.
Working backwards from 16GB: reserve about 8GB for the operating system and storage services, and at roughly 0.6GB per billion parameters at 4-bit, the largest model that fits is around 13B (Estimated).
There is no supported way to add a GPU to the F2-425: no PCIe slot and no OCuLink. Inference runs on the CPU, with Intel UHD Graphics handling media tasks, so keep to small models.
For a personal library, yes. A compact embedding model and a small chat model will run, but the F2-425 is better as the document store behind a RAG system whose model runs on stronger hardware.
Built-in networking: 2.5GbE. Adequate for a household, slower for bulk indexing.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 4GB | 4GB | None | Only configuration; 1x 4GB DDR4, 1 slot. |
The closest TerraMaster model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Business at 32 and Local LLM at 29, which points to small offices that need shared storage with private AI tools and room to grow.
Look elsewhere if you are anyone who wants to run models larger than about 8B parameters, or buyers who expect to add a graphics card later.
Other AI Capable systems closest in price.
Yes, at Q4 only. Qwen3 8B needs 5.2GB at Q4 and fits (Estimated), while Q8 needs 8.9GB and is marginal (Estimated). This is a memory calculation, not a benchmark. Memory is 4GB as shipped and 16GB at most. Gemma 3 12B is the first model that does not fit cleanly at Q4, and it is marginal (Estimated). Local LLM support is limited and Unverified.
16GB is the maximum. The F2-425 ships with 4GB of DDR4 non-ECC SODIMM in a single slot, and the memory is expandable, so you can fit more yourself. The extra memory widens the list of models that load, but the lack of a GPU path and an NPU stays the same. It does not support ECC memory.
No. The F2-425 has no PCIe slot, no USB4 ports, no Thunderbolt ports and no OCuLink, and GPU expansion is marked as not supported (Vendor confirmed). The only graphics hardware is the Intel UHD Graphics inside the CPU. It handles video transcoding, not AI model work.
No. The F2-425 has 0 M.2 slots, so there is no SSD cache or fast pool from M.2. Storage goes in the 2 drive bays. If you need M.2 for caching or container storage, look for a system that lists NVMe slots.
No. The F2-425 uses DDR4 non-ECC SODIMM, with 4GB fitted in its single slot and a 16GB maximum. Without ECC there is no error correction in memory, which matters more for business records than for a home file and photo server. The 4GB fitted is tight for containers, so an upgrade toward 16GB helps Docker apps run side by side and lets Qwen3 8B fit at Q4 (Estimated).
Yes. Docker is supported through TOS Docker Manager on TOS 7. Open WebUI as a container is Vendor confirmed, and Ollama is limited, practical only for tiny models. Virtual machines are supported as well, but with 4GB shipped and a four-core Celeron, resources are tight.
Partly. 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.
Yes. AI photos support is documented by the manufacturer. See the capability table above for the source.
Yes. Surveillance AI support is documented by the manufacturer. See the capability table above for the source.
Yes. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has 2.5GbE built in. The fastest port is 2.5GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.