Configs › gpt-oss-120b on Strix Halo: 52.9 tok/s measured
gpt-oss-120b on Strix Halo: 52.9 tok/s measured
Measured gpt-oss-120b speeds: 52.9 tok/s on Strix Halo (MXFP4, llama.cpp/vulkan). 5 configs, each linked to its source, snapshot 2026-10-16.
Snapshot 2026-10-16 · 394 configs on AI boxes and 423 on GPU cards · 123 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.
Quick answers
- How fast is gpt-oss-120b on Strix Halo?
- 52.9 tok/s single-stream decode (MXFP4, llama.cpp/vulkan), the fastest of 5 measured configs on AMD Strix Halo (Ryzen AI Max+ 395), measured by wireframeslayout, source last checked 2026-09-06. source
Measured configs for gpt-oss-120b
| Model | Hardware | Quant | Backend | Decode | Context | Trust | Source |
|---|---|---|---|---|---|---|---|
| gpt-oss-120b | Strix Halo | MXFP4 | llama.cpp/vulkan | 52.9 tok/s | 128 | Extracted, source-linked | source |
| gpt-oss-120b | Strix Halo | MXFP4 | llama.cpp/vulkan | 50.6 tok/s (prefill 725) | 512 | Structured table | source |
| gpt-oss-120b | Strix Halo | MXFP4 | llama.cpp/rocm | 50.2 tok/s | 128 | Extracted, source-linked | source |
| gpt-oss-120b | Strix Halo | MXFP4 | llama.cpp/vulkan | 31.5 tok/s | 32768 | Extracted, source-linked | source |
| gpt-oss-120b | Strix Halo | MXFP4 | llama.cpp/rocm | 17.8 tok/s | 32768 | Extracted, source-linked | source |
More: all configs · hardware · news · contributors · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.