Benchmarks run by ByteOrbital on our own hardware; written with AI assistance from our results.

Two compact machines are now serious options for running language models at home. The Minisforum MS-02 Ultra pairs an Intel Core Ultra 9 285HX with a low-profile NVIDIA card. The OneXStation i1 puts an AMD Ryzen AI Max+ 395 and its Radeon 8060S on 128 GB of memory that the CPU and GPU share. We own both, use them daily, and ran an identical suite on each.

Prefer to watch? Nova and Dex run through every chart in our six-minute video on YouTube.

How we tested

CPU

The CPU results do not give one winner. On 7-Zip the OneXStation scores 199,085 MIPS against 115,351 (73% ahead), and its decompression score is 222,087 against 106,227. It also leads single-thread sysbench (5,790 vs 4,423 events/s), x264 (314 vs 233 fps) and x265 (120 vs 99.0 fps). Yet on all-thread sysbench the MS-02 Ultra scores 146,024 against 90,108 events/s, and on Blender CPU rendering it takes 60.1 s against 64.3 s. Building the Linux kernel is a draw: 57.7 s on the OneXStation, 59.2 s on the MS-02 Ultra.

OneXStation (8060S) is 73% faster than MS-02 Ultra at 7-Zip compression

7-Zip benchmark, all threads. · MIPS · higher is better

MS-02 Ultra
115,351
OneXStation (8060S)
+73% vs MS-02 Ultra
199,085
View data table
OneXStation (8060S) is 73% faster than MS-02 Ultra at 7-Zip compression: values in MIPS
SeriesValue (MIPS)RunsMin to maxStd devVersionCommand
MS-02 Ultra115,3513112,220 to 117,0772,462p7zip-full from Ubuntu apt (7z b)7z b -mmt=24
OneXStation (8060S)199,0853198,654 to 199,526436p7zip-full from Ubuntu apt (7z b)7z b -mmt=32
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

MS-02 Ultra is 62% faster than OneXStation (8060S) at all-thread CPU speed

bo-bench sysbench multi. · events/s · higher is better

MS-02 Ultra
+62% vs OneXStation (8060S)
146,024
OneXStation (8060S)
90,108
View data table
MS-02 Ultra is 62% faster than OneXStation (8060S) at all-thread CPU speed: values in events/s
SeriesValue (events/s)RunsMin to maxStd devVersionCommand
MS-02 Ultra146,0243146,013 to 146,486270sysbench from Ubuntu aptsysbench cpu --threads=24 --time=15 run
OneXStation (8060S)90,108390,014 to 90,804431sysbench from Ubuntu aptsysbench cpu --threads=32 --time=15 run
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

Memory

This is where the unified-memory design shows. STREAM copy runs at 146,172 MB/s on the OneXStation and 57,955 MB/s on the MS-02 Ultra; triad is 107,624 against 41,596 MB/s. The MS-02's memory is 2 x 32 GB DDR5 rated at 5600 MT/s but running at 4800 MT/s, so these figures use the configured speed, and we have not tested at the rated one. The OneXStation's LPDDR5 runs at its rated 8000 MT/s. Latency is more even: at 1 GiB it is 175 ns on the MS-02 against 177 ns on the OneXStation, with the MS-02 faster in the small caches (1.02 ns against 1.46 ns at 32 KiB).

OneXStation (8060S) wins every memory bandwidth test, by 152-203%

STREAM, all threads. Copy, scale, add and triad are the four standard STREAM kernels. · GB/s · higher is better

+159% vs MS-02 Ultra OneXStation (8060S) leads in 4 of 4

050100150CopyScaleAddTriad5814636.711141.610841.6108
View data table
OneXStation (8060S) wins every memory bandwidth test, by 152-203%: values in GB/s
SeriesTestValue (GB/s)RunsMin to maxStd devVersionCommand
MS-02 UltraCopy58357.9 to 580.036stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=24 OMP_PROC_BIND=spread ./stream
MS-02 UltraScale36.7336.6 to 36.70.065stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=24 OMP_PROC_BIND=spread ./stream
MS-02 UltraAdd41.6341.5 to 41.60.055stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=24 OMP_PROC_BIND=spread ./stream
MS-02 UltraTriad41.6341.6 to 41.60.03stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=24 OMP_PROC_BIND=spread ./stream
OneXStation (8060S)Copy1463146 to 1460.029stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=32 OMP_PROC_BIND=spread ./stream
OneXStation (8060S)Scale1113111 to 1110.057stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=32 OMP_PROC_BIND=spread ./stream
OneXStation (8060S)Add1083108 to 1080.054stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=32 OMP_PROC_BIND=spread ./stream
OneXStation (8060S)Triad1083108 to 1080.031stream.c (virginia.edu, sha pinned in assets.lock.json), gcc from aptcd /var/tmp/bo-bench/stream && OMP_NUM_THREADS=32 OMP_PROC_BIND=spread ./stream
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

Storage

The drives differ, so this compares drives rather than platforms. The OneXStation has a 1 TB Acer Predator GM7; the MS-02 Ultra carries a Samsung 990 PRO 1 TB and a Crucial P310 1 TB. Sequential read is 7,226 MB/s (990 PRO), 7,064 MB/s (P310) and 6,910 MB/s (GM7). The GM7 also has the lowest queue-depth-1 read latency: 37.6 µs median against 48.4 µs for the P310 and 54.5 µs for the 990 PRO. Write results on the MS-02 varied too much while its services were running to quote fairly, so we are testing all three drives properly on their own for a separate SSD round-up.

GPU, encode and homelab

For hardware video encoding the OneXStation's VAAPI HEVC path reaches 755 fps (31.5x real time) and H.264 reaches 441 fps. On the MS-02 Ultra, NVENC HEVC does 371 fps and the Intel iGPU's VAAPI HEVC does 405 fps by default, 451 fps in our RTX-card configuration. The RTX card also does AV1 at 393 fps. We ran Blender GPU rendering only on the MS-02: 26.9 s on CUDA and 13.0 s on OptiX for the BMW scene. We have no GPU speech-to-text run for the OneXStation, but on CPU it manages 3.78x real time against 2.68x for the MS-02.

OneXStation (8060S) tops hardware video encoding: 755 fps (vaapi_hevc)

Big Buck Bunny 1080p, 60 s source. · fps · higher is better

+4.6% vs MS-02 Ultra - igpu_arc MS-02 Ultra - ada_cuda leads in 2 of 7

0200400600800nvenc_hevcintel_vaapi_h…intel_vaapi_h…nvenc_h264nvenc_av1vaapi_h264vaapi_hevc371371389414400405451427322393441755
View data table
OneXStation (8060S) tops hardware video encoding: 755 fps (vaapi_hevc): values in fps
SeriesTestValue (fps)RunsMin to maxStd devVersionCommand
MS-02 Ultranvenc_hevc3713368 to 3711.97distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -c:v hevc_nvenc -preset p5 -b:v 6M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultraintel_vaapi_h2643893387 to 4017.69distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v h264_vaapi -b:v 8M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultraintel_vaapi_hevc4053397 to 42413.6distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v hevc_vaapi -b:v 6M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - ada_cudanvenc_hevc3713369 to 3711.51distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -c:v hevc_nvenc -preset p5 -b:v 6M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - ada_cudaintel_vaapi_h2644143409 to 4258.19distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v h264_vaapi -b:v 8M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - ada_cudaintel_vaapi_hevc4513451 to 4668.73distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v hevc_vaapi -b:v 6M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - ada_cudanvenc_h2643223319 to 3221.36distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -c:v h264_nvenc -preset p5 -b:v 8M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - ada_cudanvenc_av13933392 to 3930.556distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -c:v av1_nvenc -preset p5 -b:v 5M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - igpu_arcintel_vaapi_h2644003398 to 4127.76distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v h264_vaapi -b:v 8M -f null - 2>&1 | tr '\r' '\n'
MS-02 Ultra - igpu_arcintel_vaapi_hevc4273414 to 44113.4distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v hevc_vaapi -b:v 6M -f null - 2>&1 | tr '\r' '\n'
OneXStation (8060S)vaapi_h2644413441 to 4420.435distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v h264_vaapi -b:v 8M -f null - 2>&1 | tr '\r' '\n'
OneXStation (8060S)vaapi_hevc7553754 to 7571.6distro ffmpeg (see inventory)ffmpeg -hide_banner -benchmark -y -t 60 -i /var/tmp/bo-bench/assets/bbb-clip/bbb_1080p_60s.mp4 -an -vaapi_device /dev/dri/renderD128 -vf format=nv12,hwupload -c:v hevc_vaapi -b:v 6M -f null - 2>&1 | tr '\r' '\n'
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

For a homelab, Postgres favours the MS-02 Ultra: 4,789 transactions per second against 3,078, with 3.34 ms average latency against 5.20 ms. LAN throughput is level at 2.34 and 2.35 Gbit/s.

Local LLM

The Qwen3 30B-A3B result needs care. The OneXStation generates 92.0 tok/s and processes prompts at 2,261 tok/s on its GPU. The MS-02 Ultra figures for that model (14.4 and 290 tok/s) are CPU-only, because the model needs CPU offload on the 16 GB card and we did not run that by default. The 6.4x and 7.8x ratios compare a GPU with a CPU, which is the realistic choice on that machine but not like for like.

For models that fit both GPUs the picture is tighter. On gpt-oss-20b, generation is 79.4 tok/s on the OneXStation (ROCm), against 64.3 tok/s on the MS-02's card in its default configuration and 61.2 tok/s in the dedicated card run. Qwen3 14B is 25.6 against 22.7 tok/s. Prompt processing on small models goes to the MS-02's card: Llama 3.1 8B is 1,849 tok/s on CUDA against 1,495 tok/s on ROCm (1,772 on Vulkan), and Qwen3 14B is 954 against 808 tok/s (1,010 on Vulkan).

8060S ROCm runs gpt-oss-120b MXFP4; RTX 2000E Ada can't fit it

llama.cpp llama-bench. On gpt-oss-120b MXFP4: 8060S ROCm 56.3 vs 8060S Vulkan 53.9. · tokens per second · higher is better

+3.4% vs 8060S Vulkan 8060S ROCm leads in 6 of 6

020406080100Llama 3.1 8BInstruct Q4_K_MQwen3 14B Q4_K_Mgpt-oss-20b MXFP4Qwen3 30B-A3B Q4_K_M(MoE)Llama 3.3 70BInstruct Q4_K_Mgpt-oss-120b MXFP46.8942.242.323.246.344.63.6422.719.212.825.624.89.1164.361.240.279.475.914.442.99291.15.315.2856.353.9
View data table
8060S ROCm runs gpt-oss-120b MXFP4; RTX 2000E Ada can't fit it: values in tokens per second
SeriesTestValue (tokens per second)RunsMin to maxVersionCommand
285HX CPULlama 3.1 8B Instruct Q4_K_M6.8916.89 to 6.89llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -ngl 0 -m /var/tmp/bo-bench/models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
285HX CPUQwen3 14B Q4_K_M3.6413.64 to 3.64llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -ngl 0 -m /var/tmp/bo-bench/models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
285HX CPUgpt-oss-20b MXFP49.1119.11 to 9.11llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -ngl 0 -m /var/tmp/bo-bench/models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
285HX CPUQwen3 30B-A3B Q4_K_M (MoE)14.4114.4 to 14.4llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -ngl 0 -m /var/tmp/bo-bench/models/Qwen3-30B-A3B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E AdaLlama 3.1 8B Instruct Q4_K_M42.2142.2 to 42.2llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E AdaQwen3 14B Q4_K_M22.7122.7 to 22.7llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E Adagpt-oss-20b MXFP464.3164.3 to 64.3llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E Ada - ada_cudaLlama 3.1 8B Instruct Q4_K_M42.3142.3 to 42.3llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E Ada - ada_cudaQwen3 14B Q4_K_M19.2119.2 to 19.2llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
RTX 2000E Ada - ada_cudagpt-oss-20b MXFP461.2161.2 to 61.2llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)LD_LIBRARY_PATH=/opt/cuda-12.8/lib64:/opt/bo-bench/llama.cpp-b12300-cuda/bin /opt/bo-bench/llama.cpp-b12300-cuda/bin/llama-bench -fa 1 -m /var/tmp/bo-bench/models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
Ryzen 395 CPULlama 3.1 8B Instruct Q4_K_M23.2123.2 to 23.2llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -ngl 0 -m /models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
Ryzen 395 CPUQwen3 14B Q4_K_M12.8112.8 to 12.8llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -ngl 0 -m /models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
Ryzen 395 CPUgpt-oss-20b MXFP440.2140.2 to 40.2llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -ngl 0 -m /models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
Ryzen 395 CPUQwen3 30B-A3B Q4_K_M (MoE)42.9142.9 to 42.9llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -ngl 0 -m /models/Qwen3-30B-A3B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmLlama 3.1 8B Instruct Q4_K_M46.3146.3 to 46.3llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmQwen3 14B Q4_K_M25.6125.6 to 25.6llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmgpt-oss-20b MXFP479.4179.4 to 79.4llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmQwen3 30B-A3B Q4_K_M (MoE)92192 to 92llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/Qwen3-30B-A3B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmLlama 3.3 70B Instruct Q4_K_M5.3115.31 to 5.31llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/Llama-3.3-70B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S ROCmgpt-oss-120b MXFP456.3156.3 to 56.3llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/kfd --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-10.0-strix-llama_20261102T063225 llama-bench -fa 1 -m /models/gpt-oss-120b/gpt-oss-120b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S VulkanLlama 3.1 8B Instruct Q4_K_M44.6144.6 to 44.6llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S VulkanQwen3 14B Q4_K_M24.8124.8 to 24.8llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/Qwen3-14B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S Vulkangpt-oss-20b MXFP475.9175.9 to 75.9llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/gpt-oss-20b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S VulkanQwen3 30B-A3B Q4_K_M (MoE)91.1191.1 to 91.1llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/Qwen3-30B-A3B-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S VulkanLlama 3.3 70B Instruct Q4_K_M5.2815.28 to 5.28llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/Llama-3.3-70B-Instruct-Q4_K_M.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
8060S Vulkangpt-oss-120b MXFP453.9153.9 to 53.9llama.cpp b12300 (ms01 CUDA build, sm_89); ms03 kyuz0 toolbox images pinned in hosts.yaml (~b12297)podman run --rm --device /dev/dri --group-add keep-groups --security-opt seccomp=unconfined -v /srv/models:/models:ro docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv-performance_20261101T205203 llama-bench -fa 1 -m /models/gpt-oss-120b/gpt-oss-120b-MXFP4.gguf -p 512 -n 128 -d 0,4096,16384,32768 -r 3 -o json 2>/var/tmp/bo-bench/llama.err || { tail -5 /var/tmp/bo-bench/llama.err; exit 1; }
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

Series with no bar for a model could not fit it (see what_fits in findings.json).

Big models are the OneXStation's territory: gpt-oss-120b generates at 56.3 tok/s, and Llama 3.3 70B at 5.31 tok/s with 152 tok/s of prompt processing. Context length costs speed. With 32k of context, Qwen3 30B-A3B on the OneXStation's ROCm backend still generates 39.1 tok/s.

Efficiency is closer than a headline figure suggests once you compare like for like at the wall. Llama 3.1 8B generation is 0.393 tok/s per watt on ROCm (0.402 on Vulkan) against 0.396 for the MS-02's card, which is level. On gpt-oss-20b it is 0.727 against 0.633. The OneXStation's real efficiency advantage shows on the larger models only it can run.

OneXStation (8060S) wins every tokens per watt test, by 2-558%

Generation speed divided by average wall power over the test. · tokens/s per watt · higher is better

+14% vs 8060S Vulkan 8060S ROCm leads in 3 of 6

00.20.40.60.81Llama 3.1 8BInstruct Q4_K_MQwen3 14B Q4_K_Mgpt-oss-20b MXFP4Qwen3 30B-A3B Q4_K_M(MoE)Llama 3.3 70BInstruct Q4_K_Mgpt-oss-120b MXFP40.0770.3870.3960.3930.4020.2030.0380.1940.1860.2150.2030.110.090.4370.6330.7270.6120.3360.1330.770.8750.3540.0430.0430.5310.467
View data table
OneXStation (8060S) wins every tokens per watt test, by 2-558%: values in tokens/s per watt
SeriesTestValue (tokens/s per watt)
285HX CPULlama 3.1 8B Instruct Q4_K_M0.077
285HX CPUQwen3 14B Q4_K_M0.038
285HX CPUgpt-oss-20b MXFP40.09
285HX CPUQwen3 30B-A3B Q4_K_M (MoE)0.133
RTX 2000E AdaLlama 3.1 8B Instruct Q4_K_M0.387
RTX 2000E AdaQwen3 14B Q4_K_M0.194
RTX 2000E Adagpt-oss-20b MXFP40.437
RTX 2000E Ada - ada_cudaLlama 3.1 8B Instruct Q4_K_M0.396
RTX 2000E Ada - ada_cudaQwen3 14B Q4_K_M0.186
RTX 2000E Ada - ada_cudagpt-oss-20b MXFP40.633
8060S ROCmLlama 3.1 8B Instruct Q4_K_M0.393
8060S ROCmQwen3 14B Q4_K_M0.215
8060S ROCmgpt-oss-20b MXFP40.727
8060S ROCmQwen3 30B-A3B Q4_K_M (MoE)0.77
8060S ROCmLlama 3.3 70B Instruct Q4_K_M0.043
8060S ROCmgpt-oss-120b MXFP40.531
8060S VulkanLlama 3.1 8B Instruct Q4_K_M0.402
8060S VulkanQwen3 14B Q4_K_M0.203
8060S Vulkangpt-oss-20b MXFP40.612
8060S VulkanQwen3 30B-A3B Q4_K_M (MoE)0.875
8060S VulkanLlama 3.3 70B Instruct Q4_K_M0.043
8060S Vulkangpt-oss-120b MXFP40.467
Ryzen 395 CPULlama 3.1 8B Instruct Q4_K_M0.203
Ryzen 395 CPUQwen3 14B Q4_K_M0.11
Ryzen 395 CPUgpt-oss-20b MXFP40.336
Ryzen 395 CPUQwen3 30B-A3B Q4_K_M (MoE)0.354
Caveatms01: RAM runs at 4800 MT/s, rated 5600 MT/s (2 x 32 GB DDR5). Memory bandwidth and CPU-side LLM numbers are measured at the configured speed, not the rated one.

Average power covers the whole test window including model load: compare across machines, not as a spec.

On thermals, we have temperature data only for the RTX card, which peaked at 79 °C in a sustained run. We have no comparable temperature series for the OneXStation.

What fits where

Specifications
Llama 3.1 8B Instruct Q4_K_M (4.9 GB)OneXStation: yes; MS-02 Ultra: yes (CUDA)
Qwen3 14B Q4_K_M (9.0 GB)OneXStation: yes; MS-02 Ultra: yes (CUDA)
gpt-oss-20b MXFP4 (12.1 GB)OneXStation: yes; MS-02 Ultra: yes (CUDA)
Qwen3 30B-A3B Q4_K_M (18.6 GB)OneXStation: yes; MS-02 Ultra: needs CPU offload
gpt-oss-120b MXFP4 (63.4 GB)OneXStation: yes; MS-02 Ultra: no (CPU offload only)
Llama 3.3 70B Q4_K_M (42.5 GB)OneXStation: yes; MS-02 Ultra: no (CPU offload only)

Who should buy which

We apply fixed rules: a gap of 15% or more is a clear win, and smaller gaps are judgement calls.

  • Big models (30B to 120B class) fully on the GPU: the OneXStation. It holds models up to 63 GB; the MS-02 Ultra's card tops out at 12 GB.
  • Fast chat with small models (up to about 15B): a close call. The OneXStation leads by 9% on Llama 3.1 8B (46.3 vs 42.3 tok/s). Our judgement is to decide on price, noise and size, which we did not measure.
  • Long prompts, RAG and coding agents: a simple rule picks the OneXStation (2,261 vs 290 tok/s on Qwen3 30B-A3B), but that compares its GPU with the MS-02's CPU. On models the MS-02's card holds, the card is as quick or quicker at prompt processing.
  • CPU-heavy work such as archiving: the OneXStation on 7-Zip, though the MS-02 wins all-thread sysbench and Postgres. Our judgement is to pick on your own workload.
  • Hardware video transcoding: the OneXStation, at 755 fps against 451 fps.

The MS-02 Ultra has one thing the numbers do not capture: its RTX card can be swapped, so the GPU half of this comparison is not fixed. Its memory also ran below its rating, which may change CPU-side results. Price, noise, size, warranty and availability were not measured.

Verdict

If local AI is the main job, buy the OneXStation: it runs models the MS-02 Ultra's 16 GB card cannot hold. Choose the MS-02 Ultra if you want a conventional x86 workstation with a replaceable GPU, strong all-thread CPU throughput and Postgres speed, and mostly run models under about 15B.

Some links below may earn us a commission at no cost to you. See our affiliate disclosure.

What we left out

We did not measure price, noise, size or warranty, and we did not test the MS-02 Ultra at its rated memory speed.