32GB max RAM · 8 bays · 1GbE
AINAS Score 32
No verified price

AI CapableASUSTOR
Why AI Capable: Intel Celeron N5105 (4C/4T) with Intel UHD Graphics, no NPU, no PCIe or GPU expansion and a 16GB memory ceiling; small CPU-only models via ASUSTOR's Ollama app at best.
Updated
Can the ASUSTOR Lockerstor 2 Gen2 (AS6702T) run local AI?
Only lightly. Expect 7B to 8B class models (Estimated) on the ASUSTOR Lockerstor 2 Gen2 (AS6702T): Intel Celeron N5105 does the work with up to 16GB of memory, which is fine for experiments rather than a daily assistant.
Without any GPU option, the processor carries the whole inference load.
Local LLM support is documented by the manufacturer.
The Lockerstor 2 Gen2 suits a buyer who wants a compact two-bay NAS with 2 x 2.5GbE and four M.2 slots, and who only needs light local AI as an extra. It has an Intel Celeron N5105 with four cores, 4GB of DDR4 memory that expands to 16GB, ADM 5, Docker and virtual machine support, and a three-year warranty. ASUSTOR's Ollama + Open WebUI app is vendor confirmed for this series. Photo Gallery facial recognition and Surveillance Center, with up to 44 IP cameras, are also vendor confirmed. Small models are the realistic limit: Qwen3 8B fits at Q4 (Estimated). It scores 30 on the AINAS Score, which reflects a storage-first machine.
The 16GB memory ceiling and the lack of any GPU path rule it out for buyers who want mid-size or large models. Inference is CPU-only, and Gemma 3 12B is only marginal at Q4 (Estimated), while Mistral Small 3 24B does not fit at Q4 (Estimated). There is no NPU, no listed PCIe slot and no ECC memory. Anyone planning 14B models or bigger, or fast responses, should look for a much higher memory ceiling, a PCIe slot that can take a graphics card, or an NPU. Otherwise plan on external compute for inference.
Amazon UK
4GB
£375£469
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 | Vendor confirmed | ASUSTOR App Central lists Ollama + Open WebUI (v0.11.4.r01, ADM 5.0+) for the AS67 series. CPU-only inference; 4GB base RAM and 16GB max limit model size. The app does not install a model by default to avoid failures on some NAS models. |
| Ollama | Yes | Vendor confirmed | Ollama + Open WebUI app lists AS67 as a supported series; installs via Docker. |
| Open WebUI | Yes | Vendor confirmed | Bundled with the Ollama + Open WebUI App Central package. |
| RAG | Limited | Vendor confirmed | ASUSTOR markets NAS as RAG knowledge base storage with vector DBs (Chroma, pgvector, Weaviate) via Docker; Open WebUI includes RAG. ASUSTOR recommends external compute for inference. |
| AI search | No | Unverified | No AI or semantic file search feature found in ADM 5.1 material. |
| GPU expansion | No | Vendor confirmed | No PCIe expansion slot on spec page and no iGPU. ASUSTOR states its NAS models do not feature a discrete GPU. |
| Hardware acceleration | Limited | Vendor confirmed | Intel UHD Graphics iGPU (Intel ARK) used for 4K H.265 hardware decoding/transcoding per ASUSTOR; no NPU. GPU use for AI inference not documented. |
| AI photos | Yes | Vendor confirmed | Photo Gallery 4.2 (ADM 5.1.1+) lists AS67 series; ADM 5.1 adds facial recognition grouping powered by locally run AI. |
| AI video | No | Unverified | No AI video feature found in ADM documentation. |
| Surveillance AI | Yes | Vendor confirmed | Surveillance Center 2.8.4 (app 1663) and Surveillance Center 3 (app 1661) list AS67; up to 44 IP cameras per ASUSTOR spec page licence table. |
Calculated from the Lockerstor 2 Gen2 (AS6702T)'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 Lockerstor 2 Gen2 (AS6702T)'s use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The Lockerstor 2 Gen2 is poor for local LLM work. The Ollama + Open WebUI app is vendor confirmed, but inference is CPU-only, 4GB base memory with a 16GB maximum limits model size, and Qwen3 8B fits at Q4 (Estimated).
The Lockerstor 2 Gen2 is poor for private RAG on its own. Support is limited and vendor confirmed: Chroma, pgvector and Weaviate can run in Docker containers, and Open WebUI includes RAG, but ASUSTOR recommends external compute for model inference.
Poor for AI search. No AI or semantic file search feature was found in ADM 5.1 material, and that absence is Unverified rather than confirmed, so there is no confirmed AI search route on the Lockerstor 2 Gen2 today.
Poor overall for AI photos, despite a supported feature. Photo Gallery 4.2 lists the series and ADM 5.1 adds facial recognition grouping run locally (vendor confirmed), but the Celeron N5105 has no NPU to speed up wider AI tasks.
Poor in the AINAS rating, though the software is present. Surveillance Center lists the series with up to 44 IP cameras (vendor confirmed), but the Lockerstor 2 Gen2 has no NPU to handle camera analytics.
Poor for home lab use against the AINAS rating. Docker and virtual machines are both supported, and there are four M.2 slots and 2 x 2.5GbE, but memory tops out at 16GB and no PCIe slot is listed.
Poor for business AI work. It carries a three-year warranty, but there is no ECC memory, memory tops out at 16GB, and networking is 2 x 2.5GbE only, which limits shared or heavy workloads.
Poor for creative AI. There is no GPU expansion (vendor confirmed), and hardware acceleration is limited to 4K H.265 decoding and transcoding on the Intel UHD Graphics, with GPU use for AI inference not documented.
Plan on upgrading. 4GB is fitted as standard, while the Lockerstor 2 Gen2 (AS6702T) accepts 16GB of DDR4-2933 SO-DIMM, which is the difference between only very small models of around 1B to 3B parameters and 7B to 8B class models (Estimated).
Our estimate for the biggest model that fits in 16GB is about 13B parameters at 4-bit, after setting aside 8GB for the system. Long context windows reduce that figure (Estimated).
A graphics card is not an option on the Lockerstor 2 Gen2 (AS6702T), which has neither a PCIe slot nor OCuLink. Inference runs on the CPU, with Intel UHD Graphics handling media tasks, so keep to small models.
As the storage half of a RAG setup it works well; as the model host it is limited. Small embedding models run fine on the Lockerstor 2 Gen2 (AS6702T), while the answering model needs more compute than it has.
For network access it has 2 x 2.5GbE. Enough for light use; large imports will take a while.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 4GB | 4GB | None | Diskless; 1x 4GB DDR4 pre-installed, 2 SO-DIMM slots. |
The closest ASUSTOR model by bay count and AINAS Score. Only the specifications that differ are listed.
Best fit: small offices that need shared storage with private AI tools and room to grow. That is where the Lockerstor 2 Gen2 (AS6702T) scores highest (Business 35/100), followed by Home Lab (34/100).
Less suited to anyone who wants to run models larger than about 8B parameters and buyers who expect to add a graphics card later.
Other AI Capable systems closest in price.
Yes, Qwen3 8B fits at Q4 (Estimated), needing 5.2GB. At Q8 it needs 8.9GB and is marginal (Estimated), so it loads with little headroom. This is a calculation, not a benchmark. Inference is CPU-only through ASUSTOR's Ollama + Open WebUI app, and the 4GB base memory and 16GB maximum limit model size. The app does not install a model by default.
Yes. It ships with 4GB of DDR4-2933 SO-DIMM memory and has two SO-DIMM slots. ASUSTOR's specification gives a maximum of 16GB (2x 8GB). The memory is not ECC. A 32GB module is mentioned only as unofficial in the compatibility notes, so plan around 16GB if you want a supported setup.
No. The specification page lists no PCIe expansion slot, and ASUSTOR states its NAS models do not feature a discrete GPU. The only graphics hardware is the Intel UHD Graphics, which is used for 4K H.265 decoding and transcoding. GPU use for AI inference is not documented, and there is no Thunderbolt, USB4 or OCuLink port.
No. Estimated: Mistral Small 3 24B is 14GB at Q4 and 25GB at Q8, and with a 16GB memory ceiling it does not fit at either level. Gemma 3 12B and Qwen3 14B are only marginal at Q4 (Estimated). This is a calculation, not a benchmark. A 32GB module appears in ASUSTOR's compatibility notes only as unofficial, so plan around 16GB, and look for a system with a much higher memory ceiling if 24B models matter to you.
Yes. Docker Engine is listed for this series and needs ADM 5.1.0 or later. ASUSTOR's Ollama + Open WebUI app installs via Docker, and Chroma, pgvector and Weaviate can run in containers for RAG use. VirtualBox is also listed for the series from ADM 5.1.3, so virtual machines are supported.
Yes, it has four M.2 NVMe slots alongside its two drive bays. The slots add flash storage to the system. Nothing in the listed specification says they can serve as GPU or accelerator slots, and the system has no PCIe expansion slot listed.
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.
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.
Partly. Hardware acceleration support is documented by the manufacturer. 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.