GMKtec Featured Products
GMKtec is dedicated to redefining AI PCs, gaming PCs, office mini PCs, and creative computing experiences.
GMKtec EVO-X2 AMD Ryzen™ AI Max+ 395 AI Mini PC
Choose your option
GMKtec M6 Ultra AMD Ryzen™ 5 7640HS Mini PC
Choose your option
GMKtec EVO-X3 AMD Ryzen™ AI Max+ 395 AI Mini PC
Choose your option
GMKtec G10 AMD Ryzen™ 5 3500U Mini PC
Choose your option
GMKtec K12 AMD Ryzen™ 7 H 255 Gaming Mini PC
Choose your option
GMKtec K8 Plus AMD Ryzen™ 7 8845HS Gaming Mini PC
Choose your option
GMKtec M8 AMD Ryzen™ 5 PRO 6650H Mini PC
Choose your option
GMKtec EVO-T1 Intel® Core™ Ultra 9 285H AI Mini PC
Choose your option
GMKtec EVO-X5 Pro AMD Ryzen™ AI Max+ PRO 495 AI Mini PC
Choose your option
GMKtec K17 Plus Intel® Core™ Ultra 7 258V Mini PC
Choose your option
GMKtec G11 Plus AMD Ryzen™ 5 3500U Mini PC
Choose your option
GMKtec NEO-X1 Pro AMD Ryzen™ 9 9955HX3D + NVIDIA® GeForce RTX™ 5070 MODT Gaming PC
Choose your option
GMKtec EVO-X3 AMD Ryzen™ AI Max+ 395 AI Mini PC
Choose your option
GMKtec G5S Intel® Celeron N5095 Mini PC
Choose your option
GMKtec EVO-T2 Intel® Core™ Ultra X7 358H AI Mini PC
Choose your option
GMKtec M3 Pro Intel® Core™ i5-13500H Mini PC
Choose your option
Featured in Media
GMKtec mini PCs have been featured in hands-on reviews, product roundups, and editorial coverage from respected tech publications worldwide.
GMKtec Is Recommended By
GMKtec has been widely reviewed and recommended by leading global tech media and creators, including Lon TV, ETA PRIME, Digital Foundry, ServeTheHome, Gizmochina, NASCompares, TechTablets, Iceberg Tech, and more. These trusted voices highlight GMKtec’s performance, reliability, and innovation in mini PCs and compact computing solutions.
Featured Blog
GMKtec EVO-X5 Pro Advances from Desktop AI Supercomputer to Commercial AI Computing
Following its global debut at IFA 2026 and worldwide launch on September 28, GMKtec's flagship EVO-X5 Pro expands into commercial AI with enterprise management, scalable multi-node computing, and offline 320B AI. SHENZHEN, China — October 9, 2026 — GMKtec today introduced the next stage of its EVO-X5 Pro desktop AI supercomputer at its China launch event in Shenzhen, highlighting the system's evolution from a high-performance local AI workstation into a deployable commercial AI computing platform. First unveiled globally at IFA 2026 and officially launched through GMKtec's global online channels on September 28, the EVO-X5 Pro combines the AMD Ryzen™ AI Max+ PRO 495, up to 192GB LPDDR5X-8533 unified memory, and support for fully offline LLM inference with up to 320 billion parameters. At the IFA launch, Michael Nordquist, Vice President of Client Business at AMD, joined GMKtec to highlight the companies’ long-standing collaboration and shared vision for advancing edge AI computing. The latest launch event placed a stronger emphasis on how the EVO-X5 Pro can be deployed, managed and scaled for enterprise AI workloads. From Desktop AI Supercomputer to Commercial AI Node The EVO-X5 Pro was designed from the beginning to bring large-scale AI computing out of the cloud and onto the desktop. Its latest development extends that concept into commercial environments. Rather than positioning the EVO-X5 Pro simply as a powerful mini PC, GMKtec defines it as a desktop-class AI computing node that can operate independently or become part of a larger local AI infrastructure. With support for both Windows and Linux, the system can integrate into existing development, enterprise and computing environments while keeping AI workloads and sensitive data local. This makes the EVO-X5 Pro suitable for applications where local processing, data control, computing performance and system manageability are equally important. 192GB Unified Memory for Large-Scale Local AI One of the defining features of the EVO-X5 Pro is its up to 192GB of LPDDR5X-8533 unified memory. The unified memory architecture allows CPU and GPU resources to access a shared memory pool, providing the capacity required for demanding local AI inference and GPU workloads. GMKtec originally developed the EVO-X5 Pro around a 300-billion-parameter local AI target. Through continued engineering optimization, the platform subsequently moved beyond that target and now supports fully offline LLM inference with up to 320 billion parameters, with actual compatibility and performance depending on the model, software, quantization and workload. This capability enables users to run increasingly large AI models locally without relying entirely on cloud infrastructure. The system is designed for local AI inference, AI agent workflows, model development, professional content creation, software development and other demanding workloads. Powered by AMD Ryzen AI Max+ PRO 495 At the heart of the EVO-X5 Pro is the AMD Ryzen™ AI Max+ PRO 495, combining high-performance CPU, GPU and NPU resources in a single platform. Key capabilities include: ● 16 Zen 5 CPU cores / 32 threads ● Up to 5.2GHz boost frequency ● AMD Radeon™ 8065S graphics ● 40 Compute Units with RDNA™ 3.5 architecture ● Hardware ray tracing ● AMD XDNA™ 2 NPU ● Up to 55 TOPS of dedicated AI performance The combination provides a flexible computing platform capable of handling AI inference as well as GPU computing, content creation, development, virtualization and other professional workloads. The EVO-X5 Pro also supports up to 24TB of storage, dual 10GbE networking, USB4 connectivity, 8K display output and eGPU expansion. Strengthening Collaboration with AMD The China launch also highlighted the growing strategic collaboration between GMKtec and AMD in AI computing. Zhao Hui Ji, Vice President of Marketing at AMD Greater China, delivered a keynote addressing AMD’s technology roadmap, the accelerating shift toward AI computing at the edge, and the future of the AI PC industry. His insights further highlighted the opportunities created by combining AMD’s advanced computing platforms with GMKtec’s compact desktop AI systems, particularly in local AI inference, Agentic PC applications, and commercial AI computing. The appearance of a senior AMD executive at the launch underscored the companies’ continued collaboration in bringing high-performance AI computing from the cloud and data center closer to the desktop. Built for Commercial Management and Security The PRO designation of the EVO-X5 Pro goes beyond performance. For enterprise and institutional deployments, the system integrates AMD DASH out-of-band management, allowing IT administrators to remotely manage supported systems independently of the operating system. This can include remote power control, KVM access, BIOS configuration and system recovery, making the platform suitable for environments where multiple AI nodes need to be centrally managed. The system also incorporates a dedicated dTPM 2.0 hardware security module and hardware-based fingerprint authentication, adding security layers for enterprise deployments and sensitive local AI workloads. A dedicated 10GbE management interface further separates system management from regular network traffic. Together, these capabilities are designed to make the EVO-X5 Pro easier to deploy and manage as part of a commercial AI environment. Scaling from One Node to an AI Cluster The EVO-X5 Pro can also scale beyond a single system. At the China launch event, GMKtec demonstrated two multi-node configurations designed to illustrate how compact desktop AI systems can be combined into larger local computing environments. Five EVO-X5 Pro Nodes — Up to 960GB A five-node configuration can combine five EVO-X5 Pro systems, each with up to 192GB of unified memory, providing up to 960GB of total physical memory across the cluster. The configuration is designed for larger local AI workloads, including private enterprise AI, AI agent deployment, model validation and other applications requiring more computing and memory resources than a single node can provide. Twenty Nodes — Up to 3.84TB GMKtec also demonstrated a 20-node configuration with up to 3.84TB of total physical memory. The systems can be organized as a centrally managed AI computing pool, with resources allocated across multiple users, workloads or environments. This architecture opens up applications such as AI education laboratories, university training environments, development centers and other multi-user AI computing scenarios. Rather than scaling only through a traditional data-center model, GMKtec's approach allows AI computing capacity to grow modularly by adding compact computing nodes. Local AI and Agentic Computing The EVO-X5 Pro also supports GMKtec's Claw ecosystem, which enables users to build and deploy local AI agents through a visual workflow without requiring extensive coding. Potential applications include AI-assisted productivity, content generation, private knowledge bases and other agentic workflows. By running AI workloads locally, organizations can maintain greater control over their data and computing environment while reducing reliance on public cloud infrastructure. This local-first approach reflects GMKtec's broader vision for the Agentic PC—a computer that is not simply a device for running applications, but an active AI computing platform capable of processing information, running models and supporting intelligent workflows locally. A New Stage for Desktop AI Supercomputing The EVO-X5 Pro made its global debut at IFA 2026, where it demonstrated the potential of bringing large-scale AI computing to the desktop. The system was subsequently recognized with PCWELT Best of IFA 2026 and SmartZone BEST OF IFA 2026 Innovation Award honors. The October 9 China launch adds a new dimension to that evolution by demonstrating how the same platform can extend from a single high-performance desktop AI system into a manageable and scalable commercial AI infrastructure. From 192GB unified memory and offline 320B AI to enterprise-grade management and multi-node configurations reaching 3.84TB of total physical memory, the EVO-X5 Pro represents GMKtec's next step toward making large-scale local AI computing practical at the desktop level. One node can run large models. Multiple nodes can build an AI infrastructure. Available at Official Global Website: https://tinyurl.com/d68twx75 About GMKtec GMKtec is a global technology company focused on high-performance compact computing, desktop AI supercomputing, and intelligent edge devices. The company is evolving the traditional mini PC into a new generation of powerful local AI platforms for consumers, professionals, and businesses. With an integrated approach spanning R&D, manufacturing, branding, and global distribution, GMKtec develops compact computing solutions designed to bring high-performance computing and AI capabilities closer to users. Its portfolio includes desktop AI supercomputers, mini PCs, portable graphics solutions, and consumer and commercial NAS products. GMKtec works closely with leading technology partners including AMD, Intel, and Microsoft, and continues to invest in compact high-performance computing and local AI innovation. The brand has received international recognition including the 2025 American Good Design Award Platinum and French Design Award Gold, while the EVO-X5 Pro earned PCWELT Best of IFA 2026 and SmartZone BEST OF IFA 2026 Innovation Award at IFA 2026. Today, GMKtec products are available in more than 70 countries, supporting a growing global community of users and partners. Through its focus on compact AI computing, GMKtec is working to make powerful, private, and deployable AI infrastructure accessible beyond traditional data centers and cloud environments.
Read more
GMKtec EVO-X5 Pro Officially Launched: 192GB of Memory, AMD Ryzen™ AI Max+ PRO 495, and Offline 320B AI
A new generation of Agentic PC built for large-scale local AI computing AI is moving beyond simply running applications. The next generation of computing is increasingly about understanding context, reasoning across information, generating content, and taking action. This is the vision behind the Agentic PC. Today, GMKtec officially introduces the EVO-X5 Pro, a flagship desktop AI supercomputer designed to bring large-scale AI workloads from the cloud to the desktop. Powered by the AMD Ryzen™ AI Max+ PRO 495 and equipped with up to 192GB of LPDDR5X-8533 unified memory, the EVO-X5 Pro is built for demanding local AI inference, professional workloads, and next-generation AI agents. And during the final stages of development, GMKtec achieved a new breakthrough: the EVO-X5 Pro has moved beyond its original 300B design target and now supports fully offline LLM inference with up to 320 billion parameters, with the appropriate model, software, quantization, and system configuration. From a 300B Design Target to 320B When GMKtec began developing the EVO-X5 Pro, the system was designed around a 300B-parameter local AI target. The challenge was not simply processing a large model. Running models at this scale locally requires sufficient memory capacity, memory bandwidth, GPU resources, software optimization, and sustained system performance. Through continued engineering tuning and development, GMKtec has pushed the platform beyond its original target. The EVO-X5 Pro now supports up to 320B-parameter LLMs fully offline. This development demonstrates the headroom of the system's large unified-memory architecture and opens the door to more demanding local AI workflows without requiring users to send every request to a cloud service. Actual model compatibility and performance depend on the selected model, software, quantization, workload, and system configuration. Powered by AMD Ryzen™ AI Max+ PRO 495 At the heart of the EVO-X5 Pro is the AMD Ryzen™ AI Max+ PRO 495, combining CPU, GPU, and NPU resources on a single high-performance platform. The processor features: ● 16 Zen 5 CPU cores / 32 threads ● Up to 5.2 GHz boost frequency ● AMD Radeon™ 8065S graphics ● 40 Compute Units ● RDNA™ 3.5 architecture ● Hardware ray tracing ● AMD XDNA™ 2 NPU ● Up to 55 TOPS of AI performance The result is a platform designed not only for AI inference, but also for GPU computing, content creation, development, virtualization, and other demanding desktop workloads. The IFA 2026 debut also highlighted the growing strategic relationship between GMKtec and AMD. Michael Nordquist, Vice President of Client Business at AMD, joined GMKtec's launch event in Berlin to celebrate the EVO-X5 Pro and reaffirm the companies' shared vision for the future of edge AI. The EVO-X5 Pro also received the PCWELT Best of IFA 2026 and SmartZone BEST OF IFA 2026 Innovation Award. 192GB Unified Memory: Built for Large Models One of the defining characteristics of the EVO-X5 Pro is its up to 192GB of LPDDR5X-8533 unified memory. The system provides up to 273GB/s of memory bandwidth, while allowing up to 160GB to be dynamically allocated as graphics memory. For local AI, memory capacity matters. Large language models can require substantial memory not only for model weights, but also for context, inference workloads, and other processing requirements. By combining CPU and GPU resources with a large unified memory pool, the EVO-X5 Pro is designed to provide the capacity required by large-scale local AI workloads. This architecture is central to the system's ability to move beyond its original 300B target and support up to 320B parameters locally. Designed for Private, Local AI Running AI locally changes how users can approach sensitive workloads. Instead of relying entirely on remote cloud inference, local AI can allow organizations and individuals to keep data and workloads closer to where they are being processed. The EVO-X5 Pro is designed for applications including: ● Private large language models ● Local knowledge bases ● AI coding assistants ● AI agents and multi-agent workflows ● Research and experimentation ● Professional content creation ● Enterprise AI deployments ● Offline AI inference For businesses, this can also provide greater control over data, infrastructure, and deployment environments. Enterprise-Ready by Design The EVO-X5 Pro is not designed solely as a high-performance consumer desktop. It incorporates features intended for professional and enterprise environments, including AMD DASH, a dedicated dTPM 2.0 security chip, and OS-independent out-of-band remote management. With OOB management, IT administrators can remotely monitor, maintain, and manage systems even across different operating environments. The platform has also been validated for 24/7 continuous operation, making it suitable for workloads that require sustained computing and remote administration. High-Speed Connectivity for AI Workflows Large-scale local AI often requires more than compute performance. High-speed storage, networking, external GPUs, and multiple systems can become part of the workflow. The EVO-X5 Pro provides: ● 2 × USB4 V2 — up to 80Gbps ● 2 × USB4 — up to 40Gbps ● 2 × 10GbE LAN ● eGPU support ● Multi-system clustering ● Up to 24TB storage ● Up to 8K display output This combination provides a high-bandwidth foundation for professional AI development, local data processing, high-speed storage, and scalable multi-system workflows. Designed for Sustained Performance To support demanding workloads over extended periods, the EVO-X5 Pro uses a vapor chamber and triple-fan cooling system. The system also features a CNC-machined all-metal chassis and three performance modes, allowing users to balance performance and operating requirements according to their workload. The result is a desktop platform designed not just to run AI models, but to serve as a long-term local computing workstation. Pricing and Availability The GMKtec EVO-X5 Pro is officially available from September 28, 2026, at 12:00 AM PDT. Configuration Super Early Bird Early Bird Regular Price 192GB + 2TB $6,399 $6,599 $6,799 192GB + 4TB $6,699 $6,899 $7,099 The first 30 units are available at Super Early Bird pricing and include a Z.AI 3-month Coding Plan and 1.5× GMKtec Points. Early Bird pricing will run for seven days and includes a $50 Gift Card and 1.5× GMKtec Points. Existing EVO-X2 and EVO-X3 users can also receive an exclusive $50 OFF upgrade benefit via EDM Code: X5P39550. SHOP NOW：https://tinyurl.com/d68twx75 Full Specifications Category Specification Processor AMD Ryzen™ AI Max+ PRO 495 CPU AMD Zen 5, 16 Cores / 32 Threads, Up to 5.2 GHz GPU AMD Radeon™ 8065S, RDNA™ 3.5, 40 Compute Units, Hardware Ray Tracing NPU AMD XDNA™ 2, Up to 55 TOPS Memory Up to 192GB LPDDR5X-8533 Unified Memory Memory Bandwidth Up to 273 GB/s Graphics Memory Up to 160GB Dynamically Allocated Local AI Supports Fully Offline LLMs with Up to 320B Parameters* Storage Up to 24TB USB4 V2 2 × USB4 V2, Up to 80 Gbps USB4 2 × USB4, Up to 40 Gbps Networking 2 × 10GbE LAN Security Dedicated dTPM 2.0 Remote Management AMD DASH / OS-Independent Out-of-Band (OOB) Management Enterprise Reliability Validated for 24/7 Continuous Operation Expansion eGPU Support, Multi-System Clustering Display Up to 8K Output Cooling Vapor Chamber + Triple-Fan Cooling Performance Modes 3 Performance Modes Chassis CNC-Machined All-Metal Chassis AI Platform Agentic PC / Local AI Computing * 320B-parameter support reflects GMKtec's latest development and tuning. Actual model compatibility and performance depend on the selected AI model, software, quantization, workload, and system configuration. Building the Agentic PC Era The EVO-X5 Pro represents GMKtec's next step toward the Agentic PC — a category of personal computing where AI is not simply an application running on the system, but an integral part of how the computer works. With the combination of AMD Ryzen™ AI Max+ PRO 495, up to 192GB unified memory, and support for up to 320B-parameter LLMs locally, the EVO-X5 Pro brings large-scale AI capabilities to a desktop platform. From developers and researchers to businesses and professional creators, GMKtec's goal is to make powerful local AI computing more practical, private, and accessible — and to move the desktop from simply running applications toward becoming an active AI computing platform.
Read more
GMKtec EVO-X5 Pro: The Next-Generation Agentic PC with 192GB Memory and Local 300B AI
Meet the GMKtec EVO-X5 Pro, a next-generation Agentic PC with up to 192GB unified memory, built for local AI and 300B-parameter LLMs. Subscribe for launch updates.
Read more