Benchmark result
ThinkingCap-Qwen3.8-27B-oQ4e-mtp on Apple M5 Ultra — 93.6 tok/s
Measured with omlx manual on October 1, 2026.
What the model built
Open full screen →The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited.
Reproduce this run
toolomlx manual
modelThinkingCap-Qwen3.8-27B-oQ4e-mtp (oQ4e)
context65,536
kv cachenone
effortxhigh
samplingtemperature 1 · top_p 0.95 · top_k 20 · min_p 0 · presence_penalty 0 · repetition_penalty 1
clientv0.4.85+101
Set those in omlx, then:
llm-benchmark benchmark --model "ThinkingCap-Qwen3.8-27B-oQ4e-mtp" --tool "omlx"The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.