Mini PC brand MINIX has introduced its latest product, but it is not powered by Intel or AMD chips. In fact, it is fully NVIDIA – Meet the T4000 and T5000 Generative AI Mini Workstation.
Hyper practical for popular AI workloads such as local LLM inference, generative AI workloads, and even enterprise deployments, the NVIDIA Jetson AGX Thor chip is the silicon brain behind them, with the T4000 delivering up to 1200 Sparse FP4 TFLOPs while the T5000 goes all the way up to 2070 TFLOPs.
You’re also getting up to 2560 GPU cores, fifth-gen Tensor Cores, Multi-Instance GPU support for splitting workloads efficiently, plus a dedicated PVA 3.0 engine for vision-related tasks, which altogether means these machines can comfortably handle local inference for models in the 7B to 70B parameter range.
On the other hand, they are running the Arm Neoverse-V3AE processors (12 cores on the T4000 and 14 cores on the T5000) paired with up to 128GB LPDDR5X-4266 memory for smooth execution of multiple processes, with connectivity pushed to the prosumer level with 2x 10GbE ethernet ports, Wi-Fi 6E and Bluetooth 5.3 for wireless, dual HDMI 2.1 outputs for 4K displays, plus a mix of USB ports.
Known for its compact design, MINIX is keeping that tradition without compromising on internal space, as it comes with a 1TB PCIe 4.0 SSD with more room for upgrade, while housing a twin turbo cooling system designed to keep performance stable under sustained loads.
On the software side, it ships with Ubuntu 24.04 LTS and NVIDIA JetPack 7.1 out of the box, with full support for CUDA, TensorRT, and containerized AI workflows, and more importantly, everything runs locally, which means no cloud dependency and better control over data privacy, especially for enterprise AI use cases.
