Skip to product information
New
MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
  • MINISFORUM MS-S1 MAX 64GB Local AI Pilot Edition MINISFORUM US
MINISFORUM MS-S1 MAX-P495
  • Gallery
  • Video

Free Shipping

Free Shipping

Free RMA return

Free RMA return

0% APR plan

0% APR plan

2-Year Seller Warranty

2-Year Seller Warranty

30-Day Money-Back Guarantee

30-Day Money-Back Guarantee
MINISFORUM US

MINISFORUM MS-S1 MAX-P495

$7,399.00 $9,249.00

For bulk discount inquiries, please contact elsa@minisforum.com

  • Equipped with AMD Ryzen™ AI Max+ PRO 495
  • AI Performance · Up to 131 TOPS(Overall TOPS)
  • LPDDR5x-8533MT/s | Up to 192GB
  • USB4 V2 & Dual 10GbE & WIFI 7
  • PCIE x16 Expansion | Support RAID0, RAID1
CPU:

AMD Ryzen™ AI Max+ PRO 495

Type:

192GB RAM + 2TB SSD | Estimated shipping time: Mid October

Adapter:

US

Payment Methods:

American Express Apple Pay Cartes Bancaires Diners Club Discover Google Pay JCB Maestro Mastercard PayPal Union Pay Visa Visa Electron

Trusted Delivery:

Trusted Delivery:

Free Shipping

Free Shipping

Free RMA Return

Free RMA Return

0% APR Plan

0% APR Plan

2-Year Seller Warranty

2-Year Seller Warranty

30-Day Money-Back Guarantee

30-Day Money-Back Guarantee

Compare Specs

$7,399.00$9,249.00Save $1,850.00
 Buy Now
Purchase
Overview
Specs
Reviews

MINISFORUM MS-S1 MAX-P495

More Headroom for What's Next

Powerful Intelligence, Enabling Future Possibilities

Your Best Computing Partner

MS-S1 MAX-P495 empowers AI enthusiasts to run large language models locally, build AI applications, and explore real-time, low-latency inference with more AI and memory headroom.

MS-S1 MAX-P495 enables geeks and power users to test cutting-edge AI frameworks, deploy multimodal models, and harness up to 192GB of unified memory for demanding local AI workloads.

MS-S1 MAX is designed for creative professionals, accelerating video editing, 3D rendering, AI art, and audio production with up to 192GB of unified memory, 160GB of allocatable VRAM, and multi-display output.

Ultimate Computing for
Next-Gen AI

Ultimate Computing for Next-Gen AI

APU
131 TOPS
UMA
Peak Power
Multi-Model Support

Powered by AMD Ryzen™ AI Max+ PRO 495, integrating a Zen 5 CPU, RDNA 3.5 graphics, and a next-generation NPU. With 16 cores, 32 threads, speeds up to 5.2GHz, Radeon™ 8065S Graphics, and up to 55 TOPS NPU acceleration, it delivers more compute headroom for local AI and professional workflows.

Delivers up to 131 TOPS of total AI compute, including up to 55 TOPS of dedicated NPU performance, providing more headroom for LLMs, local inference, multimodal generation, and next-generation AI workflows.

Supports up to 192GB of LPDDR5x unified memory, with up to 160GB allocatable as VRAM. A shared CPU-GPU memory pool reduces data movement and provides exceptional memory capacity for large local AI models, complex datasets, and demanding professional workflows.

160W peak power, 120W sustained load, ensuring long-term high-intensity computing stability.
Performance benchmarks for Cinebench R23, 3DMark-TimeSpy

Compatible with diverse large models and precision levels, delivering fast and stable output.

From Desktop to Rack Building an Efficient Computing Hub

Two 192GB MS-S1 MAX units form a compact AI cluster capable of running Qwen3.5-397B locally at 16 tok/s, bringing larger-model inference to a scalable desktop setup.

Supports 2U rack deployment, where multiple MS-S1 MAX-P495 units can be cascaded into a distributed cluster, creating an efficient AI computing hub for enterprises and research teams.
In real-world testing, a four-unit cluster successfully ran the DeepSeek-R1 671B Q4 large model (380GB, Q4_0 quantization), marking a breakthrough in local large-model computation for compact AI workstations.

Top-tier Thermal Solution for
Stable Computing Power

Top-tier Thermal Solution for Stable Computing Power

Aluminum Body
Solid Copper
Dual Fans
PCM Cooling
Peak 160W

Aerospace-Grade Aluminum Chassis

Aluminum chassis integrates a diamond-matrix grill for robust durability and efficient airflow, achieving a perfect synergy of superior cooling and tech-inspired aesthetics.

Copper Base & 6 Heat-Pipe Array

The pure copper heat spreader instantly captures CPU heat, which is then efficiently distributed and dissipated by the six high-performance heat pipes for dramatically lower temperatures.

Powerful Dual Turbine Fans

They generate a powerful, focused airflow to efficiently exhaust heat to prevent heat buildup.

Advanced Phase Change Material

It liquefies under heat to perfectly fill microscopic imperfections between the chip and cooler, resulting in superior thermal transfer compared to standard pastes.

120W Stable / 160W Peak PPT

With a robust 120W stable PPT and 160W Fast PPT, our cooling system confidently handles intense workloads, translating extreme power into unwavering performance, not thermal throttling.

Global-Certificatied Safe
and Efficient Power

Global-Certificatied Safe and Efficient Power

Built-in 320W high-efficiency PSU with CCC, FCC, UL, CE, UKCA and more global certifications, ensuring stable and reliable power delivery.

Engineered for Versatile Workflows
Smart Voice
Precise Recognition
Built-in AI noise-cancelling mic supports mainstream AI models, capturing voice commands with clarity.
Multiple Modes
From Desktop to Rack
Four switchable modes-Performance, Balance, Quiet, and Rack-flexibly adapt from personal desktops to enterprise clusters.
Easy Upgrade
Slide-Out Structure Design
Slide-out chassis with integrated board tray for effortless upgrades and maintenance.
Cluster Control
One-Click Start / Stop
With reserved cascade power-on header, multiple units can be powered on/off together for effortless cluster management.
Versatile Usage
Vertical or Horizontal
Supports both vertical desktop setup and horizontal rack deployment, adaptable for personal and enterprise use.

Comparison Points with MS-S1 MAX

Full Connectivity and Expansion

Full Connectivity and Expansion

I/O Ports
Dual M.2 Slots
PCIe x16

Front

Rear

1.Power Button

2. DMIC ×2

3. USB 3.2 Gen2 (10Gbps)

4. USB4 ×2 (40Gbps, Alt DP2.0, PD out 15W)

5. 3.5mm Audio Combo Jack

6.HDMI® 2.1 FRL (Up to 8K@60Hz/4K@120Hz)

7. USB 2.0 ×2

8. USB4 V2 ×2 (80Gbps, Alt DP2.0, PD out 15W)

9. USB 3.2 Gen2 (10Gbps)

10. 10GbE LAN x2 (RJ45, Realtek RTL8127)

11. Anti-Theft Lock Hole

12. Reset Hole (Clear CMOS)

13. AC INPUT

1.Power Button

2. DMIC ×2

3. USB 3.2 Gen2 (10Gbps)

4. USB4 ×2 (40Gbps, Alt DP2.0, PD out 15W)

5. 3.5mm Audio Combo Jack

Storage & Expansion

Easy Upgrade

Dual M.2 slots (PCIE 4.0 X4＋ PCIE 4.0 X1) support up to 8+8=16TB with RAID0/1; integrated slide-out board design enables effortless upgrades and maintenance.

8TB

Per Slot Maximum SSD Capacity

RAID 0/1

Flexible RAID Support

Storage & Expansion

Easy Upgrade

Dual M.2 slots (PCIE 4.0 X4＋ PCIE 4.0 X1) support up to 8+8=16TB with RAID0/1; integrated slide-out board design enables effortless upgrades and maintenance.

8TB

Per Slot Maximum SSD Capacity

RAID 0/1

Flexible RAID Support

Flexible Customization

Full-Length Expansion

Full-length PCIe x16 slot supports industry-specific expansion needs.

PCIe x16

Supports Full-length Cards

PCIe 4.0 x4

High-speed Expansion

Flexible Customization

Full-Length Expansion

Full-length PCIe x16 slot supports industry-specific expansion needs.

PCIe x16

Supports Full-length Cards

PCIe 4.0 x4

High-speed Expansion

Accessories List

Accessories List

MS-S1 MAX-P495 Specification

CPU

Processor: AMD Ryzen™ AI Max+ PRO 495（Gorgon Halo）
Total Cores: 16 Cores

Total Threads: 32 Threads
Frequency: 3.1 GHz
Max Turbo Frequency: 5.2 GHz
TDP (Performance Mode): 120W
TDP (Balanced Mode): 95W
TDP (Quiet Mode): 60W

GPU

Graphics Cards: Radeon 8065S Graphics(Internal)
Graphics Core Count: 40
Graphics Frequency: 3000 MHz

AI Engine Capabilities

AMD Ryzen™ AI
Up to 55 TOPS (NPU TOPS)
Up to 131 TOPS (Overall TOPS)

Memory

Memery Type: 256-bit LPDDR5x (Unified Memory Architecture)
Optional SKU DDR Capacity: 192GB
Max Speed Support: 8533MT/s
Max Capacity Support: 192 GB

Storage

Storage Type:

1 × M.2 2280 NVME SSD Slot (up to 8TB,PCIe 4.0 x4)
1 × M.2 2280 NVME SSD Slot (up to 8TB,PCIe 4.0 x1)

Optional SKU SSD Capacity:

0 / 1TB / 2TB / 4TB

Cooling System
*Note: Tested in the Minisforum anechoic chamber at 50 cm.

CPU Cooling Type: Six-heatpipe radiator + PCM
CPU Fan Type: 2 × 8820mm turbo fan
MAX Speed (CPU Fan): 3600 RPM
Noise (Performance Mode): Idle 37 dB, 100% load 48 dB / 50cm
Noise (Balance Mode): Idle 35 dB, 100% load 42 dB / 50cm
Noise (Quiet Mode): Idle 32 dBA, 100% load 41.6 dB / 50cm
  • 5 stars: 3 (75%)
  • 4 stars: 1 (25%)
  • 3 stars: 0 (0%)
  • 2 stars: 0 (0%)
  • 1 star: 0 (0%)
This product has no reviews of its own. The merchant grouped it with related products, so the reviews shown are from across that group.
A
Andrea Giammarchi
iGPU max VRAM 96 - DS4 requires SSD streaming

it's a lovely machine and I've recently updated to the latest BIOS hoping to solve the VRAM allocation limit that fails with AUTO or min/max values for the iGPU. If there was an option to reserve 112GB instead of 96 keeping 16GB of RAM for the CPU this machine could fly on 80/90GB GUFF models via DwarfStar 4 but unfortunately right now only the Spark, among non Apple devices, can run DeepSeek 4 Flash in VRAM, not this box.

K
Kyle Bader
Great local AI dev machine

I've been using mine for GPU initiated IO development, but also use it for general agentic development and vLLM / lama.cpp testing. Great little machine