MINIX has officially introduced its latest compact computing solutions, the T4000 and T5000 Generative AI Mini Workstations. Designed for professionals and enterprises, these powerful yet space-efficient systems are built to handle demanding generative AI workloads, including local large language model (LLM) inference, content creation, and lightweight model training—all without relying on cloud infrastructure.
Positioned as next-generation edge AI machines, the T4000 and T5000 bring high-performance computing into a compact desktop form factor, offering a practical alternative to traditional bulky server setups. They are tailored for developers, creators, and IT teams seeking secure, on-premise AI deployment with minimal latency.
Powered by NVIDIA Blackwell Architecture
At the core of both systems is NVIDIA’s latest Jetson AGX Thor platform, built on the advanced Blackwell architecture. This enables impressive AI processing capabilities directly on-device.
The T4000 delivers up to 1200 Sparse FP4 TFLOPs of AI performance, while the more powerful T5000 pushes performance further to 2070 TFLOPs. Both configurations feature between 1536 and 2560 GPU cores, along with fifth-generation Tensor Cores designed to accelerate AI workloads.
Additional features such as Multi-Instance GPU (MIG) technology allow multiple AI tasks to run concurrently with improved efficiency, while the integrated NVIDIA PVA 3.0 vision engine enhances processing for computer vision applications.
With this level of performance, the systems are capable of running local inference on large language models ranging from 7B to 70B parameters, enabling fast, private AI processing without cloud dependency.
High-Performance CPU and Memory
To complement the GPU power, the T4000 and T5000 are equipped with Arm Neoverse-V3AE processors, offering 12 cores on the T4000 and 14 cores on the T5000. These CPUs are designed to handle complex, multi-threaded workloads efficiently.
Memory is equally robust, with support for up to 128GB of LPDDR5X RAM running at 4266 MHz, delivering bandwidth of up to 273 GB/s. This allows for smooth handling of large AI models, multi-modal applications, and simultaneous processing tasks without performance bottlenecks.
Enterprise Connectivity and Flexible I/O
The mini workstations are built with professional environments in mind, offering a wide range of connectivity options. Dual 10GbE Ethernet ports provide ultra-fast networking for data-intensive workloads, while Wi-Fi 6E and Bluetooth 5.3 ensure reliable wireless connectivity.
For display and peripherals, users get dual HDMI 2.1 outputs supporting 4K at 60Hz, along with multiple USB ports, including USB 3.2 Gen 1 Type-A and Gen 2 Type-C. The system is powered via a 24V DC input with a maximum power consumption of up to 200W.
Compact Form Factor with Scalable Storage
Despite its powerful internals, the T4000 and T5000 maintain a compact footprint, measuring just 139.3 × 131 × 76.8 mm and weighing around 1.42 kg. This makes them easy to deploy in office or studio environments without requiring dedicated server racks.
Storage is handled by a 1TB PCIe 4.0 NVMe SSD, which can be upgraded up to 4TB. The system also features a durable hybrid chassis and an advanced twin turbo intercooler cooling system to ensure stable performance during sustained workloads.
Secure On-Premise AI with Full NVIDIA Ecosystem
Security and privacy are key highlights of the T4000 and T5000. All AI processing is performed locally, ensuring sensitive data and models remain on-device without being transmitted to the cloud.
The systems come pre-installed with Ubuntu 24.04 LTS and NVIDIA JetPack 7.1, providing full compatibility with CUDA, TensorRT, NVIDIA NIM, and container-based AI development workflows.
They are also built for reliability, featuring industrial-grade protections such as a 255-level programmable watchdog, ESD immunity, and the ability to operate across a wide range of environmental conditions.
Built for Professional AI Workflows
The MINIX T4000 and T5000 are designed to support a wide range of real-world AI applications. These include running private LLM-based assistants, generating AI-driven visual and multimedia content, deploying secure enterprise AI systems, and performing lightweight training tasks such as LoRA fine-tuning and model optimisation.
By combining powerful AI capabilities with a compact and efficient design, MINIX is positioning the T4000 and T5000 as versatile solutions for businesses and creators looking to bring AI processing closer to where data is generated and used.
With the growing demand for privacy-focused and low-latency AI computing, these mini workstations represent a significant step forward in making advanced AI more accessible beyond traditional data centers.