My latest video review looks at the new Minix Elite ER936-AI Mini PC. It’s powered by AMD’s Ryzen AI 9 365 processor, part of the Strix line, which gives it ten cores and a surprisingly capable built-in GPU. My review unit came configured with 32 gigabytes of DDR5 memory running at 5600 MHz and a one-terabyte NVMe SSD. You can find it over at Amazon, or direct at Minix’s online store (compensated affiliate links).
I’ve reviewed a number of Minix devices before and have generally found them to be well built, and this one fits right in with that reputation. It’s designed to be easily serviceable, with a tool-less bottom panel that lifts off for access to the internals. Underneath, you’ll find an extra NVMe slot, the Wi-Fi card, and RAM slots.
The port selection is generous for a small system. There are two 40 gigabit USB 4.0 Type-C ports that can handle Thunderbolt and video output, plus multiple 10-gigabit-per-second USB-A ports, dual 2.5-gigabit Ethernet jacks, HDMI, DisplayPort, and a headphone/mic jack placed on the front. Power comes from a 120-watt adapter, and the system draws around 100 watts under heavy load but sips about 10 to 12 watts at idle. A small fan keeps things cool; it’s audible under stress but not particularly loud, and it does a good job of maintaining thermal consistency.
The system ships with Windows 11 Pro preinstalled and activated. General performance was solid across everyday tasks like web browsing and streaming, though I noticed a few dropped frames while playing 4K60 YouTube content. The machine scored 33 on Browserbench’s Speedometer test, putting it in line with other recent AMD and Intel processors. Its graphics capabilities stood out in particular. I tested DaVinci Resolve with 4K footage and saw smooth real-time playback of transitions and a noticeable speed advantage when rendering effects compared to other mini PCs I’ve tested.
I also experimented with local AI tools to see how it handled workloads beyond typical productivity. Using AMD’s GAIA and Lemonade applications, I ran a 20-billion-parameter open-source language model that utilized the GPU. It performed significantly faster than running similar models through Olama on the same hardware which only runs through the CPU. It’s not going to rival cloud-based AI tools, but the fact that it handled local LLM this well shows how far these compact systems have come. For anyone experimenting with automation tools like N8N or other LLM setups, this configuration offers a good entry point.
Gaming performance was better than expected. In Cyberpunk 2077 at 1080p low settings, frame rates hovered around 45 to 55 frames per second, dipping slightly in more complex scenes. No Man’s Sky performed even better, staying close to 60 FPS on the ground and climbing to around 80 in space. The GPU is the limiting factor, but there’s enough CPU headroom for an external GPU setup if you wanted more performance. On 3DMark’s Time Spy benchmark, the system scored 3,504, which puts it close to older desktop GPUs like the Nvidia GTX 1060 and 1650. The stress test showed stable cooling with a 99.3 percent consistency score.
Linux support was also solid. Ubuntu detected all of the system’s components, including Wi-Fi, Ethernet, and audio, without issue. With two NVMe slots, it’s easy to set up a dual-boot configuration for Windows and Linux. This flexibility makes it suitable for home server use or software testing environments where you want both operating systems available.
After spending some time with it, the Minix Elite ER936 feels like a well-balanced small form factor computer that can handle a mix of productivity, light gaming, and local AI tasks. The design is practical, it runs efficiently, and it delivers consistent performance under load.