MINISFORUM AI X1 Pro vs MINISFORUM MS-02 Ultra

Everyday work and coding are well suited on the AI X1 Pro and on the 32 GB MS-02 Ultra. For coding the difference is scale: the X1 Pro has 12 cores and 24 threads, the MS-02 Ultra 24 cores and 24 threads.

Gaming splits them. The AI X1 Pro is workable: its Radeon 890M is fine for esports titles at 1080p, not for current AAA games at 1440p. The MS-02 Ultra is a poor fit, with Intel Graphics (4 Xe-cores) and no discrete GPU in a chassis built for work.

For a homelab, the MS-02 Ultra is well suited: two 25GbE SFP+ ports alongside 10GbE and 2.5GbE, 4 M.2 slots and room for 256 GB of memory. The AI X1 Pro is workable with 32 GB and well suited with 96 GB, and its two 2.5GbE ports allow routing or network separation on one machine.

For local AI, check the exact model release, quantization, context length, runtime overhead and memory available after the operating system. System RAM and GPU VRAM are different limits; shared-memory allocation depends on firmware, drivers and the backend. CPU inference or partial offloading can change what loads and its speed. Installed capacity alone does not establish compatibility or useful generation speed. Check the current Ollama or LM Studio hardware documentation for the exact GPU and operating system. The 32 GB MS-02 Ultra is workable, but a large model runs on its CPU, since there is no discrete GPU and no unified memory. Capacity decides which model sizes fit and bandwidth decides how fast they answer, and MINISFORUM states the bandwidth for neither machine.

As a workstation, the MS-02 Ultra is well suited thanks to its x16 card slot. The X1 Pro is workable at 32 GB and well suited at 96 GB. The 192 GB MS-02 Ultra has not been assessed for most tasks, and the barebones kit cannot be judged until you build it.

Neither machine solders its memory, and both take SO-DIMM modules. The AI X1 Pro has 2 slots and a 96 GB maximum, so the 32 GB model can grow to that ceiling while the 96 GB model ships already at it. The MS-02 Ultra has 4 slots, a 256 GB maximum and ECC support, so even the 192 GB model leaves room.

For storage, the X1 Pro's M.2 slots take drives of 2280 length, but MINISFORUM does not state their PCIe generation, so an added M.2 drive comes without a known speed. The MS-02 Ultra has twice as many, 4 M.2 slots of 2280 length, and the first two run at PCIe 4.0, where a PCIe 5.0 drive would run at PCIe 4.0 speed. The MS-02 Ultra also takes a card in its x16 slot, while the X1 Pro has 0 PCIe slot lanes.

Both carry Wi-Fi 7 and Bluetooth 5.4, with the wireless module included. The AI X1 Pro has 2 USB4 ports that carry DisplayPort, plus 1 HDMI and 1 DisplayPort output, for up to 4 displays. MINISFORUM does not state that HDMI port's version, so do not count on 4K above 60 Hz over HDMI. The MS-02 Ultra has 3 USB4 ports with DisplayPort alt mode, 4 USB-A ports and 1 HDMI 2.1 output, also for up to 4 displays, and HDMI 2.1 features still need a matching cable and monitor.

Networking is where they part ways: the X1 Pro has two ports at 2.5GbE, while the MS-02 Ultra has four, two of them 25GbE SFP+, plus 10GbE and 2.5GbE. Its 350W supply is internal, so there is no external brick to find room for.

The fair match is the AI X1 Pro with the Ryzen AI 9 HX 470, 32 GB and 1000 GB against the MS-02 Ultra with 32 GB and 1000 GB: same memory, same NVMe PCIe 4.0 storage, both on Windows 11 Pro and both from 2026. If memory capacity is the point, set the 96 GB X1 Pro with the Ryzen AI 9 HX 370 and 2000 GB against the 192 GB MS-02 Ultra with 2000 GB. The MS-02 Ultra barebones kit has no X1 Pro counterpart, and it suits buyers who already own memory and drives.

If you want serious gaming, look beyond both: each relies on integrated graphics, and neither has an OCuLink port. Before ordering either, note what else MINISFORUM does not state: the dimensions of both machines, the X1 Pro's DisplayPort version and M.2 SATA support, and the MS-02 Ultra's PCIe slot lanes and VESA mount. If any of those matters to your build, confirm it with MINISFORUM first.