🚀 Global Launch: EVO-T1 with Intel® Core™ Ultra 9 285H – Designed for the AI Age
We are pleased to announce the worldwide launch of the EVO-T1 – GMKtec's most powerful AI mini PC ever. Designed for performance-oriented creatives, edge AI developers, and demanding users who demand maximum performance in a compact system. At its heart is the revolutionary Intel® Core™ Ultra 9 285H, equipped with state-of-the-art hybrid architecture and dedicated Intel® AI Boost NPU.
💡 GMKtec EVO-T1 – Key Highlights
-
processor: Intel® Core™ Ultra 9 285H, 16 cores / 22 threads
-
AI acceleration: Total 99 TOPS (combined CPU + GPU + NPU)
-
Grafik: Intel® Arc™ 140T GPU, 77 TOPS, 2,35 GHz
-
RAM: Up to 64 GB DDR5, dual-channel, 5600–6400 MT/s
-
memory: 3× M.2 2280 PCIe 4.0×4 SSD slots, up to 12 TB
-
Expansion: Native Oculink interface for external GPU
-
Connectivity: Wi-Fi 6, Bluetooth 5.2, Dual 2.5G LAN
-
Image output: Simultaneous 4-screen output via HDMI, DP, USB4 and Type-C
-
cooling: Two fans with vapor chamber and intelligent temperature control
-
Operating system: Windows 11 Pro preinstalled
⚙️ Powered by the Ultra 9 285H – Next-generation hybrid performance
The Intel® Core™ Ultra 9 285H is more than just a CPU – it's a hybrid AI engine. Manufactured on the advanced 3nm TSMC N3B process, it offers 6 performance cores, 8 efficiency cores, and 2 power-saving efficiency cores with up to 5.4 GHz. With 24 MB L3 cache and configurable TDP up to 80 W it is ready for deep learning, 4K video editing and more.
💥 Combined AI performance
-
CPU: 9 TOPS
-
GPU: 77 TOPS
-
NPU (Intel® AI Boost): 13 TOPS
-
Total performance: 99 TOPS (Int8)
💽 Triple PCIe 4.0 SSD slots – storage without limits
With 3× M.2 2280 PCIe 4.0×4 The EVO-T1 supports up to 12 TB Ultra-fast SSD storage. Whether for model training, large datasets, or virtual environments, storage and speed are no longer bottlenecks.
🔌 Oculink – True eGPU expansion without bottlenecks
Unlike Thunderbolt, Oculink a direct PCIe Gen4×4 connection for seamless eGPU support. Ideal for power users who RTX performance for rendering, AI training or gaming – with maximum bandwidth and minimal latency – without compromise.
🌐 Seamless connectivity & display performance
-
2× 2.5G LAN ports (Realtek 8125BG) for professional network solutions
-
Wi-Fi 6 (AX200) with up to 2.4 Gbps
-
Bluetooth 5.2 for wireless accessories
-
Quadruple screen support, including HDMI 2.1 (8K@60Hz / 4K@144Hz), DP 1.4, USB4 & Type-C
❄️ Intelligent cooling with vapor chamber
The EVO-T1 has active Dual fan cooling and Vapor Chamber heat dissipationAutomatic switching between Silent Mode (35 W), Balanced Mode (54 W), and Performance Mode (80 W). Whether you're working quietly in the office or running intensive AI workloads, the EVO-T1 intelligently adapts.
🧠 Ready for local LLMs – DeepSeek 32B pre-installed
The EVO-T1 supports DeepSeek 32B local inference immediately after unpacking – with about 15 tokens/second. Ideal for developers who want to experiment with local AI models – Edge AI par excellence.
📦 Technical overview
component | specification |
---|---|
CPU | Intel® Core™ Ultra 9 285H, up to 5.4GHz |
GPU | Intel® Arc™ 140T |
NPU | Intel® AI Boost, 13 TOPS |
RAM | Dual DDR5 SO-DIMM, up to 64GB, 6400 MT/s |
Storage | 3× M.2 PCIe 4.0×4 SSD, up to 12TB |
Image output | HDMI 2.1, DP 1.4, USB4, Type-C |
AND | 2× 2.5G RJ45 |
WiFi/BT | WiFi 6 + Bluetooth 5.2 |
YOU | Windows 11 Pro |
Size/Weight | 154×151×73.6mm, 910g |
🔔 Global Launch – Available Now!
The EVO-T1 is now available worldwideExperience the combination of high-end desktop power, AI intelligence, and an ultra-compact design. Whether coding, modeling, rendering, or experimenting with LLMs, this is the ultimate AI Mini PC for 2025 and beyond.
🎁 Exclusive Early Bird Discount: 80€ OFF
An exclusive 80€ discount coupon was sent to all subscribers who between 16 June and 16 July registered for the EVO-T1.
✅ Please check your Inbox, Spam or Advertising folderto find the voucher.
This special offer was valid exclusively for pre-subscribers during the pre-launch phase and is only valid for a short time.