Minisforum AI NAS S5-304

Specifications

Maker	MMinisforum
Type	NAS
Processor	Intel Core 3 304 / Core 5 320 NAS (12 GB)
Memory (RAM)	12.0 GB
Usable for AI models	8.4 GB
Memory bandwidth	51 GB/s
FP16 compute	0.4 TFLOPS
Runs models on	CPU cores
Drive bays	5x M.2 2280 NVMe (8TB each, PCIe 4.0 x1); 128GB UFS system disk
NAS system	MinisCloud OS
Containers (Docker)	Yes
Virtual machines	Yes
Memory upgrade	LPDDR5x 7467 MT/s, soldered · up to 12.0 GB per the maker · non-ECC
Graphics card options	External GPU: 2x USB4 (40Gbps); no OCuLink; maker does not state eGPU support.
Notes	Intel Core 3 processor 304. 64-bit Intel CPU with 12GB soldered RAM and Docker, so small models (roughly 1-4B, quantized) can run on CPU. Only 5 cores, fanless at about 15W sustained, RAM cannot be upgraded and there is no GPU slot. Maker itself says 'lightweight AI workloads'.
Product page	store.minisforum.com (opens in a new tab)

Running AI on the Minisforum AI NAS S5-304

It can, with limits: expect small models only, at modest speed. Models run on its CPU cores, and this configuration leaves about 8.4 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.

Apps you can install to run models

Apps that run LLMs on a NAS, with their container images

AI features from Minisforum

Best LLMs for the Minisforum AI NAS S5-304

Quality	Model	Quant	Memory	Speed	Runs
B57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.	AGLing 3.0 Tiny (8B-A1.3B)Ant Group (inclusionAI)	Q4_K_M	6.3 GB	32.8 tok/s	Runs great
A76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.	IOK2 Horizon 7BInstitute of Foundation Models	Q4_K_M	8.0 GB	5.3 tok/s	Slowtight fit
A75Great quality (A) · 75 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.	XMiMo V2.6 Distill Qwen 9BXiaomi	Q5_K_M	8.2 GB	4.2 tok/s	Slowtight fit
B66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.	Nemotron 3.5 Bolt E4BNVIDIA	Q4_K_M	6.3 GB	10.7 tok/s	Runs well
A74Great quality (A) · 74 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.	STSenseNova Loop 9B PreviewSenseTime	Q4_K_M	7.6 GB	5.0 tok/s	Slowtight fit
A74Great quality (A) · 74 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.	DROrnith 1.5 9BDeepReinforce	Q5_K_M	8.4 GB	4.1 tok/s	Slowtight fit
A70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.	ATMax 9BAi2	Q4_K_M	7.0 GB	5.4 tok/s	Slow
A73Great quality (A) · 73 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	Qwen3.5 9BAlibaba	Q5_K_M	8.4 GB	4.1 tok/s	Slowtight fit
B58Good quality (B) · 58 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	JBMellum2.1 Thinking (12B-A2.5B)JetBrains	Q3_K_M	7.3 GB	23.0 tok/s	Runs welltight fit
B64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.	Gemma 4 E4BGoogle	Q4_K_M	6.3 GB	10.7 tok/s	Runs well

Best Image models for the Minisforum AI NAS S5-304

Quality	Model	Quant	Memory	Speed	Runs
A+89Excellent quality (A+) · 89 of 100: how capable the model is at FP8, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.	BFFLUX 3 Image KleinBlack Forest Labs	FP8	7.9 GB	8.8 min	Very slowtight fit · encoder unloaded after use
A76Great quality (A) · 76 of 100: how capable the model is at FP8, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.	BFFLUX.2 [klein] 4BBlack Forest Labs	FP8	6.2 GB	1.5 min	Slowencoder unloaded after use
A78Great quality (A) · 78 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.	Z-Image TurboAlibaba (Tongyi)	Q4 GGUF	6.1 GB	5.1 min	Very slowencoder unloaded after use
A78Great quality (A) · 78 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.	BFFLUX.2 [klein] 9BBlack Forest Labs	Q4 GGUF	8.0 GB	3.7 min	Very slowtight fit · encoder unloaded after use
A79Great quality (A) · 79 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.	BERNIE-Image TurboBaidu	Q4 GGUF	7.4 GB	6.6 min	Very slowtight fit · encoder unloaded after use
A81Great quality (A) · 81 of 100: how capable the model is at Q4 GGUF, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.	BERNIE-ImageBaidu	Q4 GGUF	7.4 GB	45.3 min	Very slowtight fit · encoder unloaded after use

Best Video models for the Minisforum AI NAS S5-304

Quality	Model	Quant	Memory	Speed	Runs
C49Fair quality (C) · 49 of 100: how capable the model is at FP8, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.	LLTX-Video 2BLightricks	FP8	6.1 GB	38.6 min	Slowencoder unloaded after use
A74Great quality (A) · 74 of 100: how capable the model is at FP8, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.	THY-OmniWeaving (8.3B, unified video)Tencent	FP8	7.8 GB	16.9 h	Very slowtight fit · CPU offload
C53Fair quality (C) · 53 of 100: how capable the model is at FP8, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.	Wan 2.1 T2V 1.3BAlibaba	FP8	7.1 GB	6.2 h	Very slowtight fit · encoder unloaded after use
C53Fair quality (C) · 53 of 100: how capable the model is at FP8, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.	Wan 2.1 VACE 1.3BAlibaba	FP8	8.0 GB	6.7 h	Very slowtight fit · encoder unloaded after use
D35Basic quality (D) · 35 of 100: how capable the model is at FP16, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.	SAAnimateDiff v3 (SD 1.5)Shanghai AI Lab	FP16	4.7 GB	32.7 min	Slow

Best Decision models for the Minisforum AI NAS S5-304

Quality	Model	Quant	Memory	Speed	Runs
A78Great quality (A) · 78 of 100: how capable the model is at F16, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.	OAOpenAI Privacy FilterOpenAI	F16	3.8 GB	77 ms	Runs great
A+86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.	JAJina Reranker v3.5Jina AI	Q8_0	2.1 GB	1.4 s	Very slow
A+90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.	Qwen3 Reranker 4BAlibaba	Q8_0	5.8 GB	9.5 s	Very slow
A80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.	FGLiNER2.5 DecideFastino	Q8_0	1.2 GB	624 ms	Slow
A84Great quality (A) · 84 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.	JAJina Reranker v3Jina AI	Q8_0	2.1 GB	1.4 s	Very slow
B68Good quality (B) · 68 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.	Llama Prompt Guard 2 22MMeta	Q8_0	0.7 GB	184 ms	Runs well
A+88Excellent quality (A+) · 88 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.	ZEzerank-2ZeroEntropy	Q8_0	5.8 GB	9.5 s	Very slow
A80Great quality (A) · 80 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.	SAFRIDA-DecisionsSber AI	Q8_0	1.7 GB	1.0 s	Slow