Get the Highlights.

GPD's all-new Mini PC features the Intel 18A Panther Lake processor. Its compact body is packed with a full range of ports, delivering a massive leap in performance for silky-smooth frame rates and stunning visuals. When paired with a GPU dock, it transforms into a high-performance desktop AI supercomputer.

GPD's all-new Mini PC line: Sleek, compact, and fully loaded with front and back ports. Mix and match anyway you want—performance pushed to the absolute limit.

Intel Core Ultra 3rd Generation Processors

 

Panther Lake GPU Die Variants

Intel® Core™ Ultra 7 356H

Intel® Core™ Ultra X7 358H

Powered by the all-new Intel 18A Panther Lake processor. Performance unleashed, graphics supercharged—prepare to play at your peak.

Next-gen tech to ignite your gaming power. Silky-smooth frame rates and explosive visuals—ready for you to dominate the virtual battlefield.

Build an AI Powerhouse with Panther Lake and eGPU. From training and debugging to full-scale deployment—experience blazing-fast AI workflows.

ꂃ ꁹ

Take a Closer Look.

Redefining the Mini PC: Unlike traditional units, GPD BOX is fully loaded with premium I/O, featuring the exclusive MCIO (PCIe Gen 5 × 8) port. With massive 512 Gbps bidirectional bandwidth and the G2 dock, it offers seamless support for the beastly RTX 5090 / 5090 D.

175 mm

134 mm

39.5 mm

SSD Cooling Air Intake

Two 2W Speaker Ports

3.5mm Headphone Jack

CPU Cooling Air Intake

USB A 10Gbps

USB4 v2.0 80 Gbps

MIC Ports

Power Button

USB A 10Gbps

DisplayPort 2.1

HDMI 2.1

MCIO 8i Port

SSD Cooling Air Exhaust

IEC 60320 C5 Power Port

SSD Cooling Air Exhaust

IEC 60320 C5 Power Port

2.5 Gigabit LAN Port

CPU Cooling Air Exhaust

USB A 10Gbps

BIOS Reset Port

DisplayPort 2.1

HDMI 2.1

2.5 Gigabit LAN Port

CPU Cooling Air Exhaust

BIOS Reset Port

GPD BOX（Ultra X7 358H）

GPD BOX（Ultra 7 356H）

ꂃ ꁹ

Intel Core Ultra 3rd Generation Processors

4-screen Multitasking: Ultra-responsive, Silky Smooth

Powered by the all-new 3rd Gen Intel® Core™ Ultra processors, the 18A process technology delivers a monumental leap in performance, pushing AI capabilities to new heights. Both the Ultra X7 358H and Ultra 7 356H feature next-gen CPU cores for smoother operation and lightning-fast responsiveness. The powerful GPU, combined with hardware-accelerated ray tracing, delivers hyper-realistic gaming visuals and fluid, detailed 3D rendering. Meanwhile, the advanced built-in neural engine provides robust computing power for a wide range of AI tasks.

Next-gen CPU

Ultra-responsive

Powers up to

4 Monitors

 

Hardware Ray Tracing

Next-Level Graphics

AI Agent Task Execution

Multimodal. Seamless. Low-Latency

  • Productivity

    Instantly handle AI document retrieval, voice transcription, and image recognition—no extra accelerator card required.

     

    With Dynamic Power Gating and fine-grained DVFS, power consumption is slashed by nearly 30%.

     

    Full Windows & Linux compatibility for Office 365, Adobe CC, and Power BI. Elevate your daily work: faster, cooler, and longer-lasting.

  • Creativity

    Stay passionate, stay driven, and dare to dream. Even professional-grade creative projects are handled with ease.

     

    For example, you can:

    Edit seamlessly with DaVinci Resolve, Lightworks, and more.

    Produce beats and mix tracks in Cubase, Studio One, and Reaper.

    Handle 3D modeling with Cinema 4D and Autodesk 3ds Max.

    Or unleash your creative vision in Adobe Photoshop.

  • Gaming

    Gaming performance is up to 77% faster than Lunar Lake, with a 50% boost in GPU power.

     

    The next-gen GPU brings a new level of intensity to your gaming. With hardware-accelerated ray tracing, lighting and reflections reach unprecedented realism, making every immersive world more vivid than ever.

  • Programming

    Leveraging Visual Studio and VS Code on Windows, IntelliSense and GitHub Copilot AI completion keep your coding speed in perfect sync with lightning-fast compilation.

     

    Leveraging Hyper-V, Windows Subsystem for Android (WSA), and multi-instance Android Emulators.

     

    Run multiple virtual devices simultaneously on a single machine. Rapidly deploy, debug, and validate apps to significantly shorten test cycles and boost development efficiency.

  • Virtual Desktops

    Leverage VMware Workstation for desktop virtualization, fully utilizing the GPD BOX's compact size, low power consumption, and powerful performance. Ideal for:

     

    1. Software Development & Cross-Platform Testing.
    2. Network Engineering & Security Labs (Home Lab).
    3. IT Administration & Hybrid Cloud Connectivity.
    4. Legacy Software Support & Sandboxed Business Environments.
    5. Education & Certification Exam Prep.

  • AI Agent

    Up to 180 TOPS of AI power for millisecond real-time inference. With AVX-512 VNNI, DLBoost, and oneAPI, it offers native compatibility with TensorFlow and PyTorch across Windows and Linux.

     

    Dynamic Power Gating and Fine-Grained DVFS slash power consumption by 30%. With SGX/TDX ensuring data integrity, we deliver a high-efficiency, low-power, and trusted computing solution for AI Agents.

     

    You can even use it to "raise your own Claw"!

Ultra X7 vs Ultra 7: Two Paths to Peak Performance.

Intel® Core™ Ultra 7 356H: Balanced iGPU, Unmatched Expandability. Unleash Pro-Level Graphics: Level Up Your Gaming with MCIO External GPU Support.

GPD BOX（Ultra X7 358H）

GPD BOX（Ultra 7 356H）

CPU & iGPU

Intel® Core™ Ultra X7 358H
16 Cores | Up to 4.8 GHz Turbo Boost
12 Xe3 Graphics Cores | 180 TOPS AI Power
15–80W Configurable TDP

Intel® Core™ Ultra 7 356H
16 Cores | Up to 4.7 GHz Turbo Boost
4 Xe3 Graphics Cores | 100 TOPS AI Power
15–80W Configurable TDP

Storage & Expansion

Memory: 32GB / 64GB LPDDR5x 8533 MT/s
Storage: 1 × M.2 PCIe Gen 5 × 4 Slot
Expansion: 1 × M.2 PCIe Gen 4 × 2 Slot
Pre-installed: 512GB – 4TB 2280 NVMe SSD

Memory: 32GB / 64GB LPDDR5x 8533 MT/s
Storage: 1 × M.2 PCIe Gen 5 × 4 Slot
Expansion: 1 × M.2 PCIe Gen 4 × 2 Slot
Pre-installed: 512GB – 4TB 2280 NVMe SSD

I/O Ports

2 × USB4 v2.0 (80Gbps)
1 × DisplayPort 2.1 (UHBR20)
1 × HDMI 2.1 (FRL)
4 × USB 3.2 Gen2 Type-A (10Gbps)
2 × 2.5G RJ45 LAN (10/100/1000/2500 Mbps)
1 × MCIO (PCIe 5.0 × 8)

2 × USB4 v2.0 (80Gbps)
1 × DisplayPort 2.1 (UHBR20)
1 × HDMI 2.1 (FRL)
4 × USB 3.2 Gen2 Type-A (10Gbps)
2 × 2.5G RJ45 LAN (10/100/1000/2500 Mbps)

Gaming Prowess

Integrated: 4-Core Xe3 Graphics | 512 Stream Processors
Performance Comparable to Radeon 780M
External Expansion: Dual-Interface GPU Support
• USB4 v2.0 (80Gbps): PCIe 5.0 × 4 | ~31.5 GB/s Bandwidth
• MCIO Exclusive: PCIe 5.0 × 8 | ~63 GB/s Bandwidth
Ultimate Desktop Power: Supports RTX 5090 / 5090D
DLSS 4.0 Support | Up to 3x Frame Gen | 200–800 FPS Performance

Integrated: 12-Core Xe3 Graphics | 1536 Stream Processors
XeSS 3 Super Sampling | Up to 4x Frame Generation
Delivers 60–120 FPS in Most AAA Titles
Expandable: External GPU via USB4 v2.0 (80Gbps)
PCIe 5.0 × 4 Shared Lane | ~31.5 GB/s Theoretical Bandwidth
Supports Up to NVIDIA GeForce RTX 5090 / 5090D
DLSS 4.0 Support | 3x Frame Gen | Up to 200–600 FPS

AI Performance

180 TOPS Int8 Peak AI Compute
Benchmark: Comparable to NVIDIA Jetson AGX Orin 64GB
At 25W TDP: Massive Efficiency Gains
• LLM Throughput: Up to 3.9x Faster
• Text-to-Speech (TTS) Speed: Up to 3.7x Faster
• Multitasking Computer Vision: Up to 5.4x Faster

100 TOPS Int8 Aggregate AI Compute
eGPU Expansion: Optimized for Multimodal Interactive Agents
Local LLM Inference, Scientific Computing & Simulation
End-to-End Workflow: From Model Training to Deployment
High-Performance Support for Data Processing & AI Training

R23 Single / Multi-Core

In terms of CPU performance, AMD’s flagship mobile processor, codenamed Strix Halo, maintains a competitive edge over Intel’s 18A-based Panther Lake processors.

Single-Core

Multi-Core

AMD Ryzen™ AI Max+ 395（120 W）

31756

2029

AMD Ryzen™ AI 9 HX 370（54 W）

23783

2053

Intel® Core™ Ultra X7 358H（65 W）☑️

21084

2096

20737

2015

Intel® Core™ Ultra 7 356H（65 W）☑️

iGPU vs. dGPU Benchmarks

The 1536-core Intel® Arc™ B390 integrated graphics edges out the mobile NVIDIA GeForce RTX 3050 in both Time Spy and Fire Strike scores.

Time Spy

Fire Strike

AMD Radeon 8060S Graphics （AI Max+ 395）

28764

11726

13376

7136

Intel® Arc™ B390 GPU（Ultra X7 358H）☑️

13143

6213

NVIDIA GeForce RTX 3050（Notebook）

9023

4431

AMD Radeon™ 890M（AI 9 HX 370）

5521

3143

Intel® Graphics（Ultra 7 356H）☑️

CPU & GPU Benchmarks

Panther Lake CPU matches Strix Halo in power, yet its 12-core Xe3 iGPU outshines the NVIDIA GeForce RTX 3050.

Intel® Core™ Ultra X7 358H ☑️

180 TOPS

AMD Ryzen AI Max+ 395

126 TOPS

Intel® Core™ Ultra 7 356H ☑️

100 TOPS

85 TOPS

Qualcomm Snapdragon X2 Elite X2E-80-100

Int8 Raw Compute Power

80 TOPS

AMD Ryzen AI 9 HX 370

Int8 Peak AI Performance

AI Performance: Panther Lake Leading the Pack.

Versatile High-Performance AI Powerhouse.

Integrated NPU and GPU power always-on Edge AI, engineered for extreme industrial environments from -40°C to 100°C.

  • AI Edge Computing

    Use Cases

    Smart Security; Industrial Inspection; Smart Home; Transportation; Medical Imaging; Clinical AI.

    Requirements

    Basic Vision (Face Detection, QR Code): 5-10 TOPS
    Mid-Level Analytics (Multi-Object Tracking, Behavior): 10-20 TOPS
    High-Level Scenarios (Multi-Channel 4K, ADAS): 20-30 TOPS

    Ultra-High Performance (City-Scale Analytics): 30-40 TOPS
     

    Configs

    GPD BOX (Ultra X7 358H) | 32GB RAM | 512GB-1TB SSD
    Dual 2.5GbE | HDMI | Multiple USB Ports
    RS-232 / RS-485 (External Support)
    Operating Temp: -20°C to 100°C (Wide Range)

    Execution

    OS: Ubuntu / Windows IoT
    AI Framework: ONNX Runtime
    Optimization: OpenVINO INT8 Quantization
    Deployment: NPU (50 TOPS) | Lightweight Agents

  • Lightweight Chat Agents

    Use Cases

    Use Case: Personal Assistants / AI Customer Service
    • Automate interactions & slash labor costs
    • 24/7 always-on operation
    • Instant response & local data privacy

    Requirements

    24/7 Smart Customer Service:
    • Low Load (1B-3B, 10-20 req/s): 20-40 TOPS
    • Mid Load (7B-14B, 30-50 req/s): 90-150 TOPS
    All-in-One Personal Assistant:
    • OpenClaw Agentic Automation
    • Gateway Mode (Remote API) or Local <7B
    • Requirement: 100-180 TOPS

    Configs

    GPD BOX (Ultra X7 358H) | 32GB-64GB RAM | 1TB SSD
    Dual 2.5GbE LAN | Multiple I/O Ports

    Execution

    OS: Ubuntu / Windows IoT
    Engine: Intel-Optimized Ollama (NPU/GPU Auto-Switch)
    Optimization: OpenVINO INT8 Quantization
    Models: Phi-3-mini (3.8B) / GPT-OSS 20B
    Integration: WeChat / Telegram / WhatsApp via Gateway

  • Multimodal AI Agents

    Use Cases

    Visual Synthesis, Multimodal Intelligence & Instant Rendering.

    Requirements

    Unleashing Multimodal Potential: Unified Image, Audio, and Mini-GPT Text Processing.

    Configs

    Configuration A: GPD BOX (358H) | 64GB RAM | 2TB SSD

    Configuration B: GPD BOX (356H) | 64GB RAM | 2TB SSD + eGPU (Min. RTX 4060 Ti)

    Execution

    Empowered by TensorRT & OneAPI Multimodal Architecture: CPU for precision scheduling, GPU for massive LLM workloads. Flowy AI onboard, pre-loaded with Stable Diffusion 3.5 and Qwen-VL-Chat-7B. With OpenVINO optimization and Arc B390 Xe Matrix Engine acceleration, you can harness ComfyUI workflows to realize an end-to-end "Text → Image → Video" creative powerhouse.

  • Local LLM Inference

    Use Cases

    Engineered for Solo Developers, Educational Showcases, and Privacy-Centric Local Inference Environments.

    Requirements

    Powering High-Concurrency API Services & Multi-Model Orchestration: Supporting Long-Context, Random-Request Workloads for SaaS-level Production Inference.

    Configs

    GPD BOX (356H) Ultimate Edition: 64GB RAM + 8TB SSD. Connected via MCIO port to GPD G2 eGPU, recommended with NVIDIA RTX 4090 or 5090 for peak performance.

    Execution

    Optimized for vLLM & Ollama: Deploy FP8 quantized models on Ubuntu/WSL2 with direct OpenAI-compatible API support. Whether via vLLM's high-throughput engine or Ollama's streamlined framework, enjoy instant, standardized AI access.

  • Scientific Simulation

    Use Cases

    Tackling Complex Frontiers: From Climate Simulation & Molecular Dynamics to CFD and Genomics.

    Requirements

    Budget-Friendly Research & Simulation: Ideal for academic institutions and SME-scale studies. A powerhouse for specialized domains with extreme Double-Precision (FP64) performance demands.

    Configs

    GPD BOX 356H | 64GB + 8TB | MCIO High-Speed Interconnect. Paired with the "Yimao Zhixia" V100 32GB SXM2 external dock via MCIO 8i to SFF-8654 8i link. Unleash 7.8 TFLOPS of FP64 power—the ultimate workstation for scientific computing and high-precision AI.

    Execution

    Empowering Scientific Frontiers on Linux: Whether on Ubuntu or specialized Linux distros, accelerate your application layer with GROMACS (Molecular Dynamics), ANSYS Fluent/OpenFOAM (CFD), WRF/CESM (Climate), and BWA/GATK (Genomics). A unified platform for mission-critical research.

Dual M.2 PCIe Gen 5 Performance

Doubled Speed, Seamless Expansion, Unlimited Throughput — Building a Data Closed-Loop to Eliminate AI Bottlenecks.

Doubled Speed

M.2 PCIe Gen 5 × 4 Flagship Slot: Supports PCIe Gen 5 x 4 NVMe M.2 SSDs, delivering blistering sequential Read/Write speeds of up to 15,000 MB/s & 14,000 MB/s. With data throughput doubled compared to the previous generation and no capacity limits, it offers a buttery-smooth "zero-latency" experience—whether you’re instantaneously loading massive 3D rendering projects or previewing 8K raw footage in real-time.

Seamless Expansion

Reserved M.2 PCIe Gen 4 × 2 Slot: Serves as a high-speed storage hub with no capacity limits. Effortlessly accommodate massive creative assets and large-scale games, transforming your GPD BOX into a pocket-sized "Personal Data Center."

Unlimited Throughput

The core of AI lies in data flow speed. With massive 8533 MT/s memory bandwidth and a PCIe Gen 5 SSD forming an ultra-fast data channel, 10-billion-parameter models load "instantly," while local AI generation speeds are boosted by nearly 50%. No more data queuing—your AI computing power stays at full throttle at all times.

Note 1: PCIe Gen 5 SSD Slot. The PCIe Gen 5 × 4 NVMe M.2 slot comes standard with a 512GB to 4TB PCIe Gen 4 × 4 NVMe M.2 SSD factory-installed. Users can manually upgrade to a PCIe Gen 5 SSD to unlock full performance.

Note 2: PCIe Gen 4 SSD Slot (Expansion). The PCIe Gen 4 × 2 NVMe M.2 slot is empty by default for user expansion. Please note that this is not a full-bandwidth PCIe Gen 4 × 4 slot; its 2-lane configuration provides throughput equivalent to PCIe Gen 3 × 4.

Ultimate Cooling

Pure Copper Heat Sink + Diagonal Airflow System: Powerful performance demands an ice-cold core. A large-area, high-conductivity pure copper plate is embedded beneath the SSD to rapidly draw away heat, while an independent diagonal turbo fan provides precision coverage of the thermal core by optimizing the airflow path. Even under continuous high-load AI computing or massive data throughput, the SSD remains at its optimal operating temperature, preventing thermal throttling and ensuring peak speeds stay persistent.

MCIO: Ultra-Bandwidth Interface, Shattering Mini PC Performance Ceilings

The world’s premier PCIe-based high-speed port, delivering 4x the bandwidth of OCuLink. Setting a new benchmark for peripheral performance, it achieves near-native gaming with only ~2% loss on the RTX 4090. One port, limitless expansion for every scenario.

Generational Leap

The world's only NUC to feature the flagship MCIO 8i port, engineered specifically for next-gen AI servers and powered by ultra-high PCIe 5.0 × 8 bandwidth. Compared to traditional OCuLink (PCIe 4.0 × 4), it delivers a 400% breakthrough in transmission bandwidth. With a unidirectional effective bandwidth of approx. 31.508 GB/s and a bidirectional effective bandwidth of approx. 63.015 GB/s, data flow now enters the era of "Instantaneous Transfer."

Flagship Resonance

Connect to the GPD G2 Graphics Dock via our exclusive MCIO 8i fiber-optic grade cable to fully unleash the savage performance of the RTX 4090 or 5090D. Shatter bandwidth bottlenecks and achieve near-native output with as low as 2% performance loss. Seamlessly transition between 'Professional Office' and 'Elite Gaming Hub'—whenever and wherever you need it.

Infinite Expansion

Connect to anything, at native speeds. Whether it's high-performance eGPU docks (such as NVIDIA V100/A100), AI accelerators, all-flash storage arrays, or 200/400 GbE network cards, the MCIO 8i port ensures seamless, direct-link integration. Your GPD BOX can instantaneously transform into a High-Performance Computing (HPC) node.

Order Redefined

Transforming how devices connect, the MCIO port utilizes server-grade physical protocols and enhanced circuitry for a reliable, high-performance link. By streamlining the device recognition and handshake process, it ensures your hardware is ready and stable from the moment you boot up. Engineered with anti-surge protection to prevent accidental crashes or compatibility errors, MCIO provides a more secure and streamlined way to harness the full power of elite peripherals.

Note: For details on the GPD G2 eGPU, please refer to: https://www.gpd.hk/gpdg2egpu

Ports Bandwidth Comparison

MCIO 8i（PCIe Gen 5 × 8）Aggregate Bidirectional Bandwidth ☑️

512 Gbps

USB4 V2.0 Aggregate Bidirectional Bandwidth ☑️

160 Gbps

Oculink（PCIe Gen 4 × 4）Aggregate Bidirectional Bandwidth

128 Gbps

USB4 V2.0 Asymmetrical Bidirectional (Tx/Rx) ☑️

120 Gbps

40 Gbps

USB4 V2.0 Symmetric Bidirectional ☑️

80 Gbps

Beyond Limits: What Else Can MCIO Connect?

Capable of transmitting various signals, including PCIe, SAS/SATA, and Ethernet, offering exceptional flexibility. Through different adapter cables and converters, the MCIO port can seamlessly connect to the following devices:

  • Dual Enterprise NVMe SSDs

    Connecting Enterprise-Grade High-Speed SSDs

     

    Directly connect to U.2/U.3 Enterprise NVMe SSDs or utilize specialized adapter cables, such as MCIO 8i (74-pin) to SlimSAS 8i, to establish a high-performance, high-reliability storage link.

    The illustration features the ICY DOCK ToughArmor MB105VP-B, a high-performance removable drive enclosure. It provides a single SlimSAS (SFF-8654) 8i port to manage two PCIe 4.0 U.2/U.3 NVMe SSDs, connected via an MCIO 8i (74-pin) to SlimSAS (SFF-8654) 8i cable. For external use, a DC power source is required to power the enclosure through a DC-to-15-pin SATA adapter cable.

  • 8-Bay Storage Enclosure

    Connecting Cost-Effective SATA/SAS Hard Drives

     

    Expand a single port into connections for multiple traditional HDDs or SSDs using a breakout cable (such as MCIO 8i to 8x SATA), providing a highly economical and scalable storage solution.

    The illustration features the ICY DOCK ToughArmor MB998SP-B, a high-performance removable drive enclosure providing eight 2.5” SATA HDD/SSD drive bays. With 8 individual SATA ports, it connects to the system via an MCIO 8i to 8x SATA breakout cable. For external use, a DC power source is required to power the enclosure through a DC-to-dual-15-pin SATA adapter cable.

  • 8K Edit & Color Grading

    MCIO 8i to PCIe 5.0 Adapter Board

     

    High-Performance Connectivity: Transform the MCIO 8i port into a PCIe slot to connect external professional Video I/O cards (e.g., Blackmagic DeckLink 8K Pro). Since 8K video previewing demands extreme throughput and ultra-low latency, the MCIO direct-link provides reference-grade 8K output to professional monitors. This achieves color depth and synchronization precision that no USB or standard HDMI port can replicate.

    The illustration demonstrates the GPD BOX connected to a dual MCIO 8i to PCIe 5.0 x16 Gen5 female adapter via the MCIO port. A Blackmagic DeckLink 8K Pro capture card is installed in the adapter's PCIe slot to handle high-demand video tasks.

     

    Connection & Power Specifications:
    •   Independent Power Supply: Both the adapter card and the capture card require an additional 6-pin independent power supply for stable operation.

    •   High Expandability: This PCIe adapter is highly versatile and can easily connect to other standard PCIe devices, such as professional sound cards and AI acceleration cards, fully unleashing the expansion potential of your system.

  • Connect Enterprise GPUs

    Fields with Extreme FP64 Performance Requirements

     

    When your GPD BOX meets the "Yimao Zhixia" V100 GPU dock via the MCIO port, performance is instantly awakened. Your slim device gains data-center-level computing power in an instant, effortlessly handling AI inference, 4K rendering, and large-scale model training. With Plug-and-Play simplicity and no more cable clutter, unleash the roaring potential of your workstation directly on your compact desktop.

    Shown here is the GPD BOX connected to an NVIDIA V100 via the MCIO port. This setup transforms the GPD BOX into a high-performance computing (HPC) workstation, delivering a level of numerical rigor and professional simulation efficiency that consumer-grade GPUs simply cannot match. In fields like scientific computing and simulation (such as CFD and structural analysis), precision is everything. The NVIDIA V100 stands out with its native FP64 double-precision units, delivering a staggering 7.8 TFLOPS of double-precision performance—4.7x that of the RTX 5090 and 6x that of the RTX 4090.

2 x USB4® Version 2.0 Ports, 80Gbps Transmission Rate

USB 80Gbps doubles the bandwidth, fully unleashing PCIe Gen 4 performance. It supports a wide range of external peripherals with high-speed stability. Note: This port is designed for data transmission and does not support power delivery (PD charging).

2 x USB4® Version 2.0 Ports

Featuring the latest USB 80Gbps protocol, it doubles the bandwidth compared to the previous USB4 generation. This allows your mobile SSDs to fully unleash the power of PCIe Gen 4 performance, delivering unprecedented data transfer speeds.

80Gbps Transmission Rate

Flexible High-Speed Bandwidth: Achieve a symmetrical 80Gbps (Send/Receive) transmission rate, or utilize the asymmetrical mode for 120 / 40Gbps (Send/Receive) throughput. This dynamic bandwidth allocation ensures optimal efficiency for high-performance displays and massive data transfers.

Peak Effective Bandwidth up to 15.754 GB/s

In symmetric mode, the effective unidirectional bandwidth is approximately 9.6 GB/s; in asymmetric mode, the effective unidirectional data bandwidth reaches approximately 14.0 to 14.5 GB/s.

Support GPD G1 / G2 eGPU Docks

Versatile Expansion: Seamlessly connect to 8K displays, portable SSDs, high-speed hubs, or the GPD G1/G2 mobile eGPU docking stations. Note: This port is dedicated to high-speed data transmission and does not support power delivery (charging) for external devices.

Note: For details on the GPD G1 eGPU, please refer to: https://www.gpd.hk/gpdg1graphics

Note: For details on the GPD G2 eGPU, please refer to: https://www.gpd.hk/gpdg2egpu

Dual-Engine Vision: 80Gbps Torrent of Light & Shadow

Full-Throttle DisplayPort™ 2.1 + HDMI 2.1 (80 Gbps / 48 Gbps) is here! Experience the ultimate power of native 8K @ 60Hz output. No more compromises—just pure, high-fidelity visual immersion.

DisplayPort™ 2.1 Bandwidth Comparison

UHBR20 (Full-Throttle) ☑️

80 Gbps

UHBR13.5 (Performance+)

54 Gbps

UHBR10 (Standard Edition)

40 Gbps

Engine 1: DisplayPort™ 2.1 UHBR20 (80Gbps)

Native "Beyond Limits" Experience

Extreme Bandwidth: Staggering 80Gbps, a massive leap.
Pure Visuals: Supports 6K @ 60Hz native lossless output.
Next-Gen Vision: 8K @ 60Hz / 5K @ 120Hz via DSC.

Exclusive Gear: Pair with DP 2.1 UHBR cable to unlock 8K.

HDMI 2.1 Bandwidth Comparison

FRL6 (Full-Throttle) ☑️

48 Gbps

32 Gbps

FRL5

24 Gbps

FRL4

Engine 2: HDMI 2.1 FRL (48Gbps)

Full-Throttle "Immersion" Experience

Flagship Protocol: Maxing Out 48Gbps FRL.
Fluid High-Refresh: 5K @ 120Hz.
Giant Screen Vision: Peak 8K @ 60Hz Quality.

Certified Gear: Use HDMI® 2.1 (48G) Cables.

18 Gbps

TMDS

Max Out Your LAN with Dual 2.5G.

Stop Gigabit bottlenecks. GPD BOX’s dual 2.5G ports blast open your data highway instantly.

Double the Speed with Dual 2.5G LAN

2500Mbps per port. Instant response for NAS backups and LAN transfers—no more waiting.

Bandwidth Aggregation Support

Combined 5.0Gbps link via Dual 2.5G LAN for massive multitasking throughput. Gaming-Grade Latency: Hardware-optimized for rock-solid, low-ping online play.

The Ultimate Soft Router

Physical WAN/LAN isolation for seamless iKuai/OpenWrt virtualization. Build the world's most powerful all-in-one router.

Isolation & Collaboration

Surf the web on one port, edit 8K video via NAS or Workstation on the other. Zero interference, maximum speed.

Redundancy & Failover

Dual-line input with auto-failover. Keep your professional streams and servers online 24/7.

Wi-Fi 6E: The Express Lane

Unlock high-speed bands and leave congestion behind. Ultra-low latency and massive bandwidth for instant 8K video and pro-gaming. Enjoy a pure, lightning-fast connection.

Bluetooth 5.3: Stable & Efficient

Stronger connection, lower latency, and less power. Whether for audio or wearables, enjoy seamless stability and extended battery life.

Quad 10Gbps USB-A: End "Data Anxiety" Forever.

Featuring four native USB 3.2 Gen 2 high-speed ports. With 10Gbps throughput per port, 1GB/s transfer speeds are now the standard. Copy massive project files or back up huge libraries in the blink of an eye. Build your ultimate high-speed data hub.

Quad USB 3.2 Gen 2 Ports

With 10Gbps throughput per port, 1GB/s transfer speeds are now the standard. From massive project files to vast media libraries, backups happen in the blink of an eye. Build your ultimate high-speed data hub.

Expand Your Gear with Ease

Massive extensibility. Simultaneously connect high-speed drive enclosures, pro-gaming cams, 10GbE adapters, and external sound cards—all with dedicated bandwidth for every device.

Full-Speed Bandwidth

Experience the true power of your portable SSD. No more enduring the sluggish transfer speeds of traditional USB 3.0.

Seamless 4K Video Workflow

Four 10Gbps USB ports provide the raw power for real-time 4K editing and multi-device collaboration.

All-Dimensional Sound. Capture Inspiration as It Happens

Built-in Dual 2W Stereo Speakers. High-sensitivity microphone and a versatile 3.5mm 2-in-1 audio jack included.

Immersive Audio. Built-in Dual 2W Stereo

Experience clear, native sound for entertainment or remote learning. No external speakers required—declutter your workspace instantly.

Clear Voice, Unbound

High-gain omnidirectional microphone built-in. Experience precise voice capture in complex settings without the need for a headset.

Versatile Audio. CTIA Standard Ready

Precision support for Apple and standard headsets. High Signal-to-Noise Ratio (SNR) reveals every subtle nuance in your audio.

Immersive Audio: Supports DTS:X Ultra

Breaking hardware limits and redefining soundstage boundaries. Through professional DTS tuning and algorithmic optimization, whether using headphones or speakers, it delivers rich, layered 3D surround sound. From the cinematic grandeur of movies to precision audio positioning in gaming, experience a truly immersive audio feast.

Dual Fans, Quad Heat Pipes: Born to Stay Cool.

Unleashed by Panther Lake’s exceptional efficiency. Even under heavy workloads, it stays effortless—delivering blazing-fast speeds without breaking a sweat.

Dual Fans Unleashed: The Storm is Coming

Independent cooling for CPU and SSD. Dual high-pressure fans with opposing airflow double the intake, flushing out core heat in an instant.

Quad Heat Pipes: Instant Thermal Transfer

Dual 6mm + Dual 8mm pure copper heat pipes contact the core directly, flashing heat to massive fins. Thermal efficiency boosted by over 60%.

Full 80W Unleashed

Squeeze every last drop of performance out of Intel Panther Lake.

Smart Thermal Control: Cool to the Touch

Responds to load changes in milliseconds. Even after 24/7 operation, the chassis stays refreshingly cool.

Built-in Compact 160W GaN Power Supply.

At just 17.5 × 13.4 cm, it houses a lightweight 160W GaN power supply. Ultra-compact, highly efficient, and low-heat—delivering massive power in a tiny frame. GPD BOX is a true master of space utilization.

Ditch the Power Brick: Built-in 160W GaN

Within its ultra-slim 17.5 × 13.4 × 3.95 cm body, we’ve achieved the impossible by integrating a powerful 160W GaN power supply. Say goodbye to bulky external adapters—one simple cable is all you need to unleash peak performance.

GaN Technology: Tiny Size, 96% Efficiency

Leveraging advanced GaN semiconductor technology, this integrated power supply shrinks in size while achieving over 96% conversion efficiency. Compared to traditional silicon-based PSUs, it runs cooler and steadier under heavy loads—constantly fueling the Intel® Core™ Ultra 3 with a full 160W of power.

Ultimate Space Mastery. Monster Performance

GPD BOX is the ultimate master of space utilization. Through a precision-engineered internal layout, we’ve achieved the perfect balance of dual SSD slots and flagship cooling—all while making the 160W power supply "disappear." Minimalist desk, mobile office, plug-and-play: this is a total revolution of what "Mini" means.

Aluminum Alloy + PC Body

Only 0.94kg: A Desktop in Your Pocket

Don’t let its size fool you. Inside this 17.5cm chassis lies power that shames many bulky desktops. Powered by Intel® Core™ Ultra and 8533 MT/s high-speed memory, with integrated graphics that rival dedicated cards. From office tasks and video editing to AAA gaming, it delivers true desktop-level performance with ease.

0.94 kg

5.5–6.8 kg

Integrated Control Center: Secure Login & Smart Indicators

Support for Windows Hello and Linux login via the power button. Say goodbye to password anxiety—fast, secure, and always one step ahead.

Fingerprint Unlock

Deeply integrated fingerprint recognition that fully supports Windows Hello and offers native compatibility with Linux fingerprint unlocking. No matter your workflow, a single touch provides instant access, eliminating the need for tedious passwords.

RGB Backlight

Move beyond the cold, clinical feel of traditional hardware. With subtle shifts between white and red lighting effects—Constant, Breathing, or Transition—the light syncs with power, sleep, and hibernation modes. Let your GPD BOX breathe with you, creating a truly immersive and rhythmic tech workspace.

Smart Status Lighting

Intuitive three-mode logic at a glance:

* Steady White: System running
* Breathing White: Resting in Standby
* Steady Red: Fully powered off or Hibernating