128GB max RAM · 2 x M.2 · 6 x 10GbE
AINAS Score 34
No verified price
Traditional NASPromise
Why Traditional NAS: Intel Core i5-8500 with no NPU, no documented GPU path and a 64GB memory ceiling; Promise documents no Docker or container support, so the no Docker rule makes it traditional.
Updated
Can the Promise PegasusPro R8 run local AI?
No. The Promise PegasusPro R8 is a storage-first Traditional NAS: Intel Core i5-8500 and at most 64GB of memory leave no room for a local language model.
It is here as a reference point against systems that can run AI.
The PegasusPro R8 suits buyers who want an 8 bay storage appliance for business file work or media editing, not local AI. It has four 10GbE ports and two 1GbE ports, plus two Thunderbolt 3 ports for direct connection to a workstation. The Intel Core i5-8500 has 6 cores. The system comes with 32GB of DDR4 memory, which can be expanded to 64GB. A 350W power supply and a 3 year warranty sit alongside that. Promise positions it as a hybrid NAS and Thunderbolt DAS, managed through PROMISE Utility Pro on a PROMISE appliance OS whose name is not published. It scores 35 on the AINAS Score, which reflects its traditional class. It is built as a storage box first, and the AI side is not what it was designed to do.
The limitation that matters most is that nothing on the PegasusPro R8 can run model software. Promise documents no Docker or container support, and the local LLM row is marked no and Unverified. There is no NPU, and its 2 PCIe slots have no stated type and no documented GPU use. Memory tops out at 64GB of DDR4, ECC is not stated, and there are no M.2 slots for fast cache. Buyers who also want AI later should look for a system with documented container support, a confirmed GPU path or an NPU, and a memory ceiling above 64GB. Buyers who only want Thunderbolt and 10GbE file storage for editing are not affected, since that is what Promise built it for.
No verified price
We have not verified a listing for the Promise PegasusPro R8 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 | No | Unverified | No Docker, container or LLM support in the datasheet. |
| Ollama | No | Unverified | No evidence found. |
| Open WebUI | No | Unverified | No evidence found. |
| RAG | No | Unverified | No evidence found. |
| AI search | No | Unverified | No AI search feature documented. |
| GPU expansion | No | Unverified | PCIe slots listed but no GPU use documented. |
| Hardware acceleration | No | Unverified | No acceleration documented. |
| AI photos | No | Unverified | No feature documented. |
| AI video | No | Unverified | Application list covers editing software only; no AI video feature. |
| Surveillance AI | No | Unverified | No feature documented. |
Calculated from the PegasusPro R8'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 | Q8 | Yes | 38GB | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | Q8 | Yes | 44GB | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | Q8 | Yes | 47GB | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | Q8 | Yes | 50GB | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | Q8 | Yes | 50GB | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | Q4 | Yes | 60GB | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | None fits | No | 89GB at Q4 | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | Q3 | Yes | 54GB | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | Q2 | Yes | 62GB | Not measured | Estimated |
One verdict per workload, from the PegasusPro R8's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The PegasusPro R8 is poor for local LLM work. The datasheet shows no Docker, container or LLM support (Unverified), so although 64GB of memory would hold Llama 3.3 70B at Q4 (Estimated), nothing on the system can run it.
Private RAG is a poor match here. RAG, Open WebUI and Ollama are all marked as unsupported and Unverified, with no evidence found, and the PegasusPro R8 has no documented container support to host any of them.
AI search is also poor on the PegasusPro R8. The datasheet documents no AI search feature (Unverified), and with no NPU and no GPU path there is no hardware on the system aimed at that kind of indexing.
Expect no AI photo features from the PegasusPro R8, so the verdict is poor. No such feature is documented (Unverified), and the system has no NPU or GPU to support face or object recognition.
Surveillance is poor on this system. No surveillance feature is documented (Unverified), and the PegasusPro R8 lists no NPU or GPU, so there is no documented way to analyse camera footage on the box itself.
We cannot confirm how well the PegasusPro R8 suits a home lab. Docker and virtual machine support are not documented either way, though it offers 8 bays, 6 cores and memory expandable to 64GB.
Business use is an acceptable fit for the PegasusPro R8. It has 8 bays, four 10GbE ports, ECC not stated, a 3 year warranty and PROMISE Utility Pro for management, which suits file storage rather than AI.
Creative AI is a poor fit. GPU expansion and hardware acceleration are both marked unsupported and Unverified, and PCIe slots are listed with no GPU use documented, so the Core i5-8500 does all the work.
Out of the box the PegasusPro R8 has 32GB; the official ceiling is 64GB of DDR4. A memory upgrade is the cheapest way to make it more useful for AI.
Working backwards from 64GB: 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 93B (Estimated).
There is no supported way to add a GPU to the PegasusPro R8: no PCIe slot and no OCuLink. Inference runs on the CPU, so keep to small models.
Not well. A useful private RAG setup needs more memory or compute than the PegasusPro R8 offers.
Built-in networking: 4 x 10GbE + 2 x 1GbE. Plenty for a shared knowledge base with several users.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| Standard, 32GB | 32GB | None |
The closest Promise model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Home Lab at 50 and Business at 49, which points to home lab owners who want one box for storage, containers, media and a few AI services.
Look elsewhere if you are buyers who expect to add a graphics card later.
Other Traditional NAS systems closest in overall score.
No. The datasheet documents no Docker, container or LLM support (Unverified), so nothing on the PegasusPro R8 can run model software. Memory alone would not help either. Llama 4 Scout 17Bx16E needs 67GB at Q4, which is more than the 64GB memory ceiling (Estimated). Even the 32GB it ships with would not hold it. Buyers who want local models should look for container support, a GPU path or an NPU, and more memory headroom.
Yes. The PegasusPro R8 ships with 32GB of DDR4 and the listed maximum is 64GB. The memory is long-DIMM DDR4, and the system is marked as expandable rather than soldered. ECC is not stated, so do not assume error-corrected modules are supported. More memory will help caching and file serving, but it does not add any AI capability, because no container or LLM support is documented.
No. The datasheet lists no M.2 or NVMe slots, so the PegasusPro R8 has 0 NVMe slots. Storage is through its 8 bays. That means no dedicated flash cache drive on the board. Buyers who want an M.2 cache or a fast boot and application volume should look for a system that lists M.2 or NVMe slots in its specification.
No GPU use is documented, so treat it as unsupported. The PegasusPro R8 lists 2 PCIe slots but does not state their type, and it is unclear whether the 10GbE card uses one. GPU expansion is marked unsupported and Unverified. There is also no OCuLink port. Buyers who need a graphics card should look for a documented GPU path, such as a confirmed full-length PCIe slot.
No, as far as Promise documents. The datasheet lists no Docker or container support, and the local LLM and Ollama rows are both marked no and Unverified. The PegasusPro R8 runs a PROMISE appliance OS, whose name is not published, managed through PROMISE Utility Pro, and virtual machine support is not stated either. That is why it sits in the traditional class: its 6-core Intel Core i5-8500 and 32GB of memory serve file storage, not apps.
It has 2 Thunderbolt 3 ports and 0 USB4 ports. The Thunderbolt 3 ports let the PegasusPro R8 work as direct attached storage for a workstation, which is why Promise describes it as a hybrid NAS and Thunderbolt DAS. For network use it also has four 10GbE ports and two 1GbE ports. Buyers who need USB4 specifically should look elsewhere.
No. Ollama support is not yet verified. See the capability table above for the source.
No. Open WebUI support is not yet verified. See the capability table above for the source.
No. AI photos support is not yet verified. See the capability table above for the source.
No. Surveillance AI support is not yet verified. See the capability table above for the source.
No. Hardware acceleration support is not yet verified. See the capability table above for the source.
It has 4 x 10GbE + 2 x 1GbE built in, and the 2 PCIe slots listed; type not stated can take a faster network card. The fastest port is 4 x 10GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.