Benchmark result
Qwen3.8-27B-MTPLX-Optimized-Quality on Apple M5 Max — 44.8 tok/s
Measured with oMLX 0.7.0rc1 on September 25, 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; the frame only adds a line that tells this page how tall it is.
How this run compares
1st fastest of 4 runs of this model on Apple M5 Max · median 40.0 tok/s.
sort
- 144.8tok/s83.4oMLX · KV nonethis run
- 244.1tok/s82.2oMLX · KV none
- 335.8tok/s89.3oMLX · KV none
- 432.8tok/s88.9oMLX · KV none
Reproduce this run
tooloMLX 0.7.0rc1
modelQwen3.8-27B-MTPLX-Optimized-Quality
context65,536
kv cachenone
effortmedium
think limitnone
samplingtemperature 0.6 · top_p 0.95 · top_k 20 · min_p 0 · presence_penalty 0 · repetition_penalty 1
clientv0.4.80+98
Set those in oMLX, then:
llm-benchmark benchmark --model "Qwen3.8-27B-MTPLX-Optimized-Quality" --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.