4GB max RAM · 2 bays · 1GbE
AINAS Score 22
£159.99£219.99
Amazon UK · Verified

Traditional NASSynologyDiscontinued
Why Traditional NAS: Intel Celeron J4125 (4C/4T) with UHD 600 iGPU but no NPU, no PCIe/OCuLink GPU path, and only 6 GB maximum memory (under the 8 GB threshold), so only tiny CPU-only models are practical.
Updated
Can the Synology DS423+ run local AI?
No. The Synology DS423+ is a storage-first Traditional NAS: Intel Celeron J4125 and at most 6GB of memory leave no room for a local language model.
It is here as a reference point against systems that can run AI.
The DS423+ suits buyers who want a 4-bay Synology for file storage, Surveillance Station and Synology Photos, not for local AI. It has 4 bays and 2 M.2 slots, and an Intel Celeron J4125 with 4 cores. Surveillance Station is vendor confirmed for up to 40 IP cameras, with 2 free licences. It runs Synology DSM with Docker and virtual machine support, and it ships with 2GB of DDR4 non-ECC memory. The AINAS Score is 20, which reflects how little room it leaves for AI work.
The memory ceiling rules it out for anyone who wants to run language models. The DS423+ tops out at 6GB, and it has no NPU, no PCIe slot, no OCuLink and no USB4, so there is no GPU path either. Estimated: even Llama 3.2 3B does not fit at Q4 or Q8. Buyers who want local LLMs, private RAG or creative AI should look for a system with at least 32GB of expandable memory, an NPU or a GPU path such as a PCIe slot or OCuLink, and 2.5GbE or faster networking. Synology has also discontinued it, with its status given as Phase Out.
No verified price
We have not verified a listing for the Synology DS423+ 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 | Community confirmed | CPU-only via Ollama in Container Manager; 6 GB maximum RAM limits it to very small (about 1B to 3B) models. |
| Ollama | Limited | Community confirmed | No Ollama package; runs as ollama/ollama container in Container Manager. 6 GB RAM cap is tight. |
| Open WebUI | Limited | Community confirmed | Open WebUI container documented on Synology DSM with Ollama; running both within 6 GB RAM is tight. |
| RAG | Limited | Unverified | No Synology local RAG on this model. AI Console bridges to cloud AI providers or OpenAI-compatible endpoints (which could be self-hosted elsewhere). Open WebUI document chat via Docker plausible but not verified on this unit. |
| AI search | No | Vendor confirmed | AI Console 2.0 beta GPU-powered AI Search for Synology Drive is limited to FS3420, FS6420 and 26-series rack models; this model is not listed. |
| GPU expansion | No | Vendor confirmed | No PCIe slot, OCuLink or USB4; no GPU path. |
| Hardware acceleration | Limited | Community confirmed | J4125 has Intel UHD Graphics 600 with Quick Sync (Intel ARK); reviewer tested Plex hardware transcoding with mixed 4K results. No NPU; no AI acceleration usable for LLMs. |
| AI photos | Limited | Vendor confirmed | Synology Photos face and object recognition supported; object recognition needs memory upgraded to at least 4 GB (ships with 2 GB). |
| AI video | Limited | Vendor confirmed | No on-NAS deep video analytics (DVA series only). Surveillance Station can use on-camera AI people/vehicle detection from Synology BC500/TC500 cameras. |
| Surveillance AI | Yes | Vendor confirmed | Surveillance Station supported, up to 40 IP cameras (1080p H.264, 800 FPS); 2 free licences, more require purchase. |
Calculated from the DS423+'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 | None fits | No | 11GB at Q4 | 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 DS423+'s use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The DS423+ is a poor choice for local LLMs. Ollama only runs as a container in Container Manager, which is community confirmed as limited, and the 6GB maximum memory restricts it to very small models.
Private RAG is a poor fit on the DS423+. Synology offers no local RAG here, and document chat through Open WebUI in Docker is unverified, so treat it as unconfirmed on this system.
AI search is a poor fit and is not supported. Synology's GPU-powered AI Search for Drive in the AI Console 2.0 beta does not list the DS423+, which is vendor confirmed.
AI photos is rated poor, though partly possible. Synology Photos supports face and object recognition (vendor confirmed), but object recognition needs the memory raised to at least 4GB from the 2GB it ships with.
Surveillance is rated poor for AI work, even though Surveillance Station itself is supported. It handles up to 40 IP cameras, but there is no on-NAS deep video analytics, only on-camera detection from Synology BC500 or TC500 cameras.
Home lab use is rated poor. Docker and virtual machines are both supported, but 2GB of memory as shipped, a 6GB ceiling and a 4-core Celeron J4125 leave little room for several services.
Business use is rated poor. It has a 3 year warranty and 4 bays, but only 2 x 1GbE, non-ECC memory and discontinued status count against it for shared or growing workloads.
Creative AI is a poor fit. There is no GPU path (vendor confirmed), no NPU, and the Intel UHD Graphics 600 gives mixed Plex 4K transcoding results with no usable AI acceleration for LLMs.
Out of the box the DS423+ has 2GB; the official ceiling is 6GB of DDR4 non-ECC. A memory upgrade is the cheapest way to make it more useful for AI.
There is no supported way to add a GPU to the DS423+: no PCIe slot and no OCuLink. Inference runs on the CPU, with Intel UHD Graphics 600 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 DS423+ is better as the document store behind a RAG system whose model runs on stronger hardware.
Built-in networking: 2 x 1GbE. Adequate for a household, slower for bulk indexing.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| DS423+ 2GB | 2GB | None | 2 GB DDR4 non-ECC pre-installed plus 1 SODIMM slot; max 6 GB (2 GB + 4 GB D4NESO-2666-4G). |
The closest Synology model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Business at 23 and Home Lab at 22, 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 Traditional NAS systems closest in overall score.
No. Estimated: Llama 3.2 3B needs 2GB at Q4 and 3.4GB at Q8, and it does not fit at either level on the DS423+. The system has 2GB installed and a 6GB maximum, with the operating system and containers sharing that memory. Community notes say Ollama in Container Manager is limited to very small models of roughly 1B to 3B.
6GB is the maximum. The DS423+ comes with 2GB of DDR4 non-ECC memory and has one SODIMM slot free. The listed upgrade is a 4GB module, which brings the total to 6GB. Object recognition in Synology Photos needs at least 4GB, so the upgrade matters for that feature.
No. The DS423+ uses DDR4 non-ECC memory, so it has no error correction. The memory can be expanded, up to 6GB in total, but any module you add will also be non-ECC. Buyers who need ECC for data integrity should look for a system that lists ECC memory in its specification.
No. Synology's specification lists no PCIe slot, no OCuLink port and no USB4, and the GPU expansion capability is vendor confirmed as no. The only graphics hardware is the integrated Intel UHD Graphics 600, which is used for Plex transcoding, not for running language models.
Yes, 2. The DS423+ has 2 M.2 NVMe slots alongside its 4 drive bays. They are listed for SSDs in the manufacturer's accessory list. They do not add a GPU path or extra memory, so they do not change what the system can do for AI.
Yes. Container Manager runs Docker on Synology DSM, and virtual machines are supported too. Ollama has no package, but it runs as the ollama/ollama container, and Open WebUI is documented alongside it, both community confirmed as limited. With 2GB fitted and a 6GB maximum, running both is tight, and community notes put the ceiling at very small models of about 1B to 3B. The 4-core Celeron J4125 suits light containers for file and media services better.
Partly. Ollama support is confirmed by community reports. See the capability table above for the source.
Partly. Open WebUI support is confirmed by community reports. 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.
Yes. Surveillance AI support is documented by the manufacturer. See the capability table above for the source.
Partly. Hardware acceleration support is confirmed by community reports. See the capability table above for the source.
It has 2 x 1GbE built in. The fastest port is 1GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.