Products · IceWhale · NAS
IceWhale ZimaCube
- Processor
- Intel N100 / N95
- Memory (RAM)
- 8, 16 or 32 GB
- Launch price
- $649
- Released
- 2024
Configurations
Pick one to see which models it runs. More memory lets larger models fit; bandwidth sets how fast they answer.
| Configuration | Memory | Usable for AI | Bandwidth | LLMs that run well | Launch price |
|---|---|---|---|---|---|
| Intel N100 / N95 | 8.0 GB | 5.2 GB | 38 GB/s | 147 of 622 | $649 |
| Intel N100 / N95 | 16.0 GB | 12.0 GB | 38 GB/s | 169 of 622 | — |
| Intel N100 / N95 · community-tested upgrade, above the maker's stated maximum | 32.0 GB | 26.2 GB | 38 GB/s | 200 of 622 | — |
Specifications 8.0 GB configuration
| Maker | IceWhale |
|---|---|
| Type | NAS |
| Processor | Intel N100 / N95 NAS (8 GB) |
| Memory (RAM) | 8.0 GB |
| Usable for AI models | 5.2 GB |
| Memory bandwidth | 38 GB/s |
| FP16 compute | 0.2 TFLOPS |
| Runs models on | CPU cores |
| Drive bays | 6x 3.5/2.5in SATA + 4x M.2 NVMe (7th bay, 800 MB/s) |
| NAS system | ZimaOS; other OSes can be installed |
| Containers (Docker) | Yes |
| Virtual machines | Yes |
| Memory upgrade | DDR4 3200 SO-DIMM · 1 slot · up to 16.0 GB per the maker · 32.0 GB reported working by owners · non-ECC |
| Graphics card options | PCIe: 1x PCIe 3.0 slot (maker lists it as 'Gen 3.0, 2 GB/s x1'; lane count not stated). Half-height cards up to…. External GPU: Zima's 800W OCuLink GPU Docking Station ($179.90, PCIe 4.0 x4, GPUs up to 650 W) is listed as compatible with…. Maker-supported graphics cards: NVIDIA Tesla P4 (8 GB, 50 W); NVIDIA GTX 1650 LP (4 GB, 75 W); NVIDIA RTX 3050 LP (6 GB, 70 W). |
| Notes | Intel Processor N100. N100 with 8 GB shipped: CPU-only small models. The maker recommends i5-1235U or better for the Intel build of ZimaOS AI. |
| Product page | www.zimaspace.com (opens in a new tab) |
Running AI on the IceWhale ZimaCube
It can, with limits: expect small models only, at modest speed. Models run on its CPU cores, and this configuration leaves about 5.2 GB for them after the NAS system and its storage services. The grades below assume the NAS isn’t busy with backups or media work at the same time. More memory is the upgrade that matters most: see the configurations.
Apps you can install to run models
| App | How it installs |
|---|---|
| Ollama | From the NAS app store |
| Open WebUI | From the NAS app store |
| AnythingLLM | From the NAS app store |
AI features from IceWhale
| Feature | What it does | Where the AI runs |
|---|---|---|
| ZimaOS AI (AI search in Files) | Optional module installed with 'zpkg install zimaos_ai' (NVIDIA) or 'zpkg install zimaos_ai_intel' (Intel iGPU); downloads a local model and adds natural-language/semantic search of files. | On the NAS |
Best LLMs for the IceWhale ZimaCube 8.0 GB
| Quality | Model | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. | Gemma 4 E2B | Q4_K_M | 4.2 GB | 13.7 tok/s | Runs well |
| 50Fair quality (C) · 50 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. | Ling 3.0 Tiny (8B-A1.3B) | Q3_K_M | 5.1 GB | 28.4 tok/s | Runs welltight fit |
| 41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. | Unlimited OCR 2 (3B-A0.6B) | Q4_K_M | 3.6 GB | 44.5 tok/s | Runs great |
| 57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. | Nemotron Labs Dory 1.0 4B | Q3_K_M | 3.4 GB | 10.2 tok/s | Runs well |
| 38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. | Unlimited OCR (3B-A0.6B) | Q4_K_M | 3.6 GB | 44.5 tok/s | Runs great |
| 57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. | MiniCPM5 2B | Q4_K_M | 2.9 GB | 12.5 tok/s | Runs well |
| 49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. | Nemotron 3.5 Bolt E4B | Q2_K | 4.4 GB | 11.5 tok/s | Runs well |
| 36Basic quality (D) · 36 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. | Granite 4.0 H Tiny (7B-A1B) | Q3_K_M | 4.5 GB | 35.9 tok/s | Runs welltight fit |
| 47Fair quality (C) · 47 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. | Nanbeige4.2 3B | Q2_K | 4.2 GB | 12.4 tok/s | Runs well |
| 47Fair quality (C) · 47 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. | Gemma 4 E4B | Q2_K | 4.4 GB | 11.5 tok/s | Runs well |
Best Image models for the IceWhale ZimaCube 8.0 GB
| Quality | Model | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 90Excellent quality (A+) · 90 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. | FLUX 3 Image Klein | Q8 GGUF | 4.3 GB | 29.1 min | Very slowCPU offload |
| 72Great quality (A) · 72 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. | FLUX.2 [klein] 4B | Q4 GGUF | 4.6 GB | 4.3 min | Very slowtight fit · encoder unloaded after use |
| 39Basic quality (D) · 39 of 100: how capable the model is at FP8, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking. | SDXL Turbo | FP8 | 4.2 GB | 28 s | Runs well |
| 82Great quality (A) · 82 of 100: how capable the model is at FP8, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking. | Z-Image Turbo | FP8 | 4.4 GB | 10.0 min | Very slowCPU offload |
| 81Great quality (A) · 81 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking. | ERNIE-Image | Q4 GGUF | 4.1 GB | 2.0 h | Very slowCPU offload |
| 81Great quality (A) · 81 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking. | MOSS-Image SFT | Q8 GGUF | 4.1 GB | 71.4 min | Very slowCPU offload |
Best Video models for the IceWhale ZimaCube 8.0 GB
| Quality | Model | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 35Basic quality (D) · 35 of 100: how capable the model is at FP16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking. | AnimateDiff v3 (SD 1.5) | FP16 | 4.7 GB | 68.2 min | Very slowtight fit |
| 49Fair quality (C) · 49 of 100: how capable the model is at FP8, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking. | LTX-Video 2B | FP8 | 4.0 GB | 1.7 h | Very slowCPU offload |
| 44Fair quality (C) · 44 of 100: how capable the model is at FP8, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking. | CogVideoX 2B | FP8 | 4.7 GB | 4.0 h | Very slowtight fit · CPU offload |
Best Decision models for the IceWhale ZimaCube 8.0 GB
| Quality | Model | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking. | OpenAI Privacy Filter | Q8_0 | 2.4 GB | 197 ms | Runs well |
| 86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. | Jina Reranker v3.5 | Q8_0 | 1.9 GB | 4.3 s | Very slow |
| 87Excellent quality (A+) · 87 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. | Qwen3 Reranker 4B | Q5_K_M | 4.2 GB | 28.7 s | Very slow |
| 84Great quality (A) · 84 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. | Jina Reranker v3 | Q8_0 | 1.9 GB | 4.3 s | Very slow |
| 80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking. | GLiNER2.5 Decide | Q8_0 | 1.2 GB | 1.9 s | Very slow |
| 85Excellent quality (A+) · 85 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking. | zerank-2 | Q5_K_M | 4.2 GB | 28.7 s | Very slow |
| 85Excellent quality (A+) · 85 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking. | Shieldstral 1.0 3B | Q6_K | 4.4 GB | 27.6 s | Very slow |
| 80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. | FRIDA-Decisions | Q8_0 | 1.7 GB | 3.2 s | Very slow |
Other products with the Intel N100 / N95 NAS (8 GB)
K100Product page (opens in a new tab)
N150-K / N355 Mini-ITX NAS motherboard (6x SATA, 2x i226-V)Product page (opens in a new tab)
N8 6-bay NAS motherboard (Intel N100 / N150; sold elsewhere as MW-N100-NAS class board)Product page (opens in a new tab)
M1SProduct page (opens in a new tab)
P5-4M 6-bay Pocket NAS (N100 / N150 / i3-N305 / Core 3 N355)Product page (opens in a new tab)
LincStation S1Product page (opens in a new tab)
F2-424Product page (opens in a new tab)
F4-424Product page (opens in a new tab)
F6-424Product page (opens in a new tab)
F8 SSDProduct page (opens in a new tab)
F4 SSDProduct page (opens in a new tab)
F2-425 PlusProduct page (opens in a new tab)
Spec sources
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.