Ultra-Slim 3.5L Half-Height Desktop Chassis, Uncompromised Professional Performance
StarMang GT3 is an ultra-compact AI edge workstation built for space-limited desktop deployment. With a tiny 3.5L half-height form factor, it delivers professional-grade computing without bulky tower size. It supports Intel 12th~14th Gen Core i3/i5/i7 and AMD Ryzen 7000 / Ryzen 9000 CPUs, fitting perfectly in offices, labs and industrial edge terminals.
Wide NVIDIA SFF GPU Compatibility with Powerful Peak AI Compute
This mini workstation works with NVIDIA Pro 4000 SFF, Pro2000, A2000 and T1000 discrete GPUs. Max 24GB VRAM and up to 770 TOPS theoretical peak AI performance. Different GPU options cover tasks from lightweight 3D modeling, industrial design to heavy local LLM inference, data analysis and AI image/video generation.
High Expandable Memory & PCIe4.0 High-Speed Storage
GT3 features dual SO-DIMM DDR5 slots, expandable up to 96GB RAM to run multiple AI models concurrently. Dual M.2 2280 NVMe SSD bays with PCIe 4.0 standard accelerate large model loading and massive dataset read-write, ensuring smooth multitasking for local AI workloads.
Rich I/O Interfaces for Stable Professional Workflow
Comprehensive port layout: 1×HDMI,1×DP, 1×USB2.0,2×USB3.2,2×USB2.0, 1×2.5G RJ45 Ethernet, Audio Jack and dedicated CLR_CMOS quick reset button. Multi-display output and low-latency 2.5G network support multi-screen work and stable edge computing connection.
Local Offline AI Model Deployment for Multiple Industry Use Cases
It supports local deployment of 7B–35B LLMs, SD/FLUX/Wan 540P/720P text-to-image & video models, industrial CAD simulation, OCR and real-time TTS. Ideal for law firms & government sensitive-data knowledge bases, university AI training & LoRA fine-tuning, industrial edge computing, e-commerce lightweight rendering and design creation.