16GB max RAM · 6 x M.2 · 5GbE
AINAS Score 26
No verified price

AI CapableSynologyDiscontinued
Why AI Capable: AMD Ryzen R1600 (2C/4T) with no NPU or iGPU, 32 GB max DDR4 ECC, and only a PCIe Gen3 x2 network upgrade slot (E10G22-T1-Mini), so no GPU path; CPU-only small models via Container Manager.
Updated
Can the Synology DS723+ run local AI?
The Synology DS723+ can run small local models, with limits. Its AMD Ryzen R1600 and 32GB memory ceiling suit models up to the 14B class, with 32B class models a tight fit (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 confirmed by community reports.
The DS723+ suits a buyer who wants a two-bay Synology for file storage, backups and a few containers, and who treats local AI as a side task. It has 2 bays, 2 M.2 slots, an AMD Ryzen R1600 with 2 cores, and DDR4 ECC SODIMM memory that goes from the 2GB fitted up to 32GB. Docker and virtual machines are both supported. Surveillance Station handles up to 40 IP cameras, with 2 licences included. Ollama and Open WebUI run in Container Manager on the CPU (Community confirmed), once memory is added. It carries the AINAS Score of 27 and a 3 year warranty.
The main problem is that there is no route to a GPU. The one PCIe Gen3 x2 slot only takes a 10GbE network module, and the R1600 has no integrated graphics and no NPU. All AI work falls on two CPU cores, so replies from small quantised models are slow. Anyone who wants responsive local LLMs, AI search or creative AI should look for a system with a GPU expansion path or an integrated GPU, an NPU, and a higher memory ceiling. Synology has also discontinued the DS723+, which matters to anyone who wants years of vendor support.
No verified price
We have not verified a listing for the Synology DS723+ 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 | With its 2GB default the DS723+ cannot load a language model; fitted with the full 32GB, a two-core R1600 runs small quantised models in Docker slowly. |
| Ollama | Yes | Community confirmed | No Synology package; the ollama/ollama container runs under Container Manager on this two-bay x86 model, CPU only, once memory is added. |
| Open WebUI | Yes | Community confirmed | Open WebUI runs as a second container next to Ollama in Container Manager; on two cores expect slow replies. |
| RAG | Limited | Unverified | Synology ships no local RAG for the DS723+. AI Console only forwards prompts to external AI services; document chat through Open WebUI is possible but untested here. |
| AI search | No | Vendor confirmed | The DS723+ is not on the list for Synology's GPU-based AI Search, which is limited to FS3420, FS6420 and 26-series rack units. |
| GPU expansion | No | Vendor confirmed | The single PCIe Gen3 x2 slot is reserved for the E10G22-T1-Mini 10GbE card, so there is no route to a graphics card. |
| Hardware acceleration | No | Vendor confirmed | The R1600 in this unit has no integrated graphics and no NPU, so video transcoding and AI work fall on the two CPU cores. |
| AI photos | Limited | Vendor confirmed | Face recognition works in Synology Photos; object recognition needs at least 4GB, so upgrade from the 2GB it ships with. |
| AI video | Limited | Vendor confirmed | No deep video analytics on the DS723+ itself; AI detection only comes from Synology BC500 or TC500 cameras. |
| Surveillance AI | Yes | Vendor confirmed | Surveillance Station handles up to 40 IP cameras on the DS723+, with 2 licences included and more sold separately. |
Calculated from the DS723+'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 | Q8 | Yes | 19GB | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | Q8 | Yes | 19GB | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | Q8 | Yes | 24GB | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | Q8 | Yes | 28GB | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | Q8 | Yes | 28GB | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | Q4 | Yes | 25GB | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | Q4 | Yes | 29GB | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | Q4 | Yes | 31GB | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | Q4 | Yes | 32GB | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | Q4 | Yes | 32GB | 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 DS723+'s use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The DS723+ is a poor choice for local LLMs. Support is limited (Community confirmed): the 2GB default cannot load a model, and with 32GB fitted the two-core R1600 runs small quantised models slowly in Docker.
Private RAG is a poor fit here. Synology offers no local RAG for the DS723+, and document chat through Open WebUI is possible but untested, so that support is Unverified and unconfirmed.
AI search is a poor fit, because the DS723+ cannot do it. Synology's GPU-based AI Search does not list this model, and that is Vendor confirmed.
AI photos is a poor fit, with only partial support. Face recognition works in Synology Photos, but object recognition needs at least 4GB, so the 2GB it ships with must be upgraded (Vendor confirmed).
Surveillance is rated poor for AI use, even though basic recording works. Surveillance Station handles up to 40 IP cameras with 2 licences included, but the DS723+ has no deep video analytics of its own, so AI detection depends on specific Synology cameras.
Home lab use is rated poor. Docker and virtual machines are both supported, but the two-core R1600 and its network slot limit what you can run, and there is no GPU path.
Business use is rated poor for AI workloads. ECC memory is supported and the warranty is 3 years, but the DS723+ is discontinued and has only 2 x 1GbE ports built in.
Creative AI is a poor fit. There is no GPU expansion (Vendor confirmed), no hardware acceleration, and the only PCIe slot is reserved for a 10GbE module.
The DS723+ ships with 2GB of DDR4 ECC SODIMM and supports up to 32GB with ECC. Upgrading to the maximum lifts the model ceiling from only very small models of around 1B to 3B parameters to models up to the 14B class, with 32B class models a tight fit (both Estimated).
As a rule of thumb, a 4-bit model needs about 0.6GB per billion parameters plus room for context. After leaving around 8GB for the NAS operating system, 32GB leaves space for a model of roughly 40B parameters at most (Estimated).
ECC memory support is a plus for always-on systems that also hold important files.
No. The DS723+ has no PCIe slot or OCuLink port for a graphics card. Inference runs on the CPU, so keep to small models.
With limits. The DS723+ 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 1GbE. Fine for one or two users; indexing a large archive over the network will be slower.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| DS723+ 2GB ECC | 2GB | None | 2 GB DDR4 ECC SODIMM (1x 2 GB), 2 slots, max 32 GB (16 GB x 2). |
The closest Synology model by bay count and AINAS Score. Only the specifications that differ are listed.
The DS723+ scores best for Business (29/100) and Local LLM (28/100). It suits small offices that need shared storage with private AI tools and room to grow.
It is a weaker fit for buyers who expect to add a graphics card later and for teams moving large files over the network every day.
Other AI Capable systems closest in overall score.
Yes, on paper. The estimated fit shows DeepSeek-R1 32B (Qwen distill) fitting at Q4 with the full 32GB fitted (Estimated), but not at Q8 (Estimated). This is a memory calculation, not a benchmark. With only the 2GB it ships with, it cannot load a model. Even with 32GB, the two-core R1600 has no GPU help, so expect slow output.
32GB is the maximum. It ships with 2GB of DDR4 ECC SODIMM and has 2 slots, so you can fit two 16GB modules. Memory is expandable. More memory matters for AI work here, because the 2GB default cannot load a language model, and object recognition in Synology Photos needs at least 4GB.
Yes. The DS723+ uses DDR4 ECC SODIMM, and the 2GB module fitted is ECC. Memory is expandable to 32GB across 2 slots. ECC helps protect stored data from memory errors, which suits a NAS used for backups and file storage.
No. The single PCIe Gen3 x2 slot is reserved for the E10G22-T1-Mini 10GbE module, so there is no route to a graphics card (Vendor confirmed). There is also no OCuLink, Thunderbolt or USB4 port. Look for a system with a real GPU expansion path if you need one.
Yes, two. The DS723+ has 2 NVMe slots beside its 2 drive bays, which gives a compact system room for flash storage without giving up a bay. They do not add a GPU path: the only PCIe slot is a Gen3 x2 network upgrade slot for the E10G22-T1-Mini 10GbE module, and there is no OCuLink, Thunderbolt or USB4. USB is USB 3.2 Gen 1 only, and a 65W adapter powers the whole system.
Yes, both. It is on Synology's Container Manager list and its Virtual Machine Manager list. Ollama and Open WebUI run as containers on the CPU (Community confirmed), though replies on two cores are slow. Add memory first.
Yes. Ollama support is confirmed by community reports. See the capability table above for the source.
Yes. 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.
No. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has 2 x 1GbE built in, and the 1x PCIe Gen3 x2 network upgrade slot (E10G22-T1-Mini 10GbE RJ-45 module only) can take a faster network card. 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.