Benchmark result
Qwen3.6-27B-MTPLX-Optimized-Speed on Apple M5 Max — 17.4 tok/s
Measured with oMLX 0.31.3 on July 14, 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
4th fastest of 5 runs of this model on Apple M5 Max · median 19.5 tok/s.
sort
- 131.2tok/s85.6oMLX
- 221.9tok/s84.3oMLX
- 319.5tok/s58.3oMLX
- 417.4tok/s57.2oMLXthis run
- 57.4tok/s77.1oMLX
Reproduce this run
tooloMLX 0.31.3
modelQwen3.6-27B-MTPLX-Optimized-Speed
context65,536
clientv0.4.26+97
Set those in oMLX, then:
llm-benchmark benchmark --model "Qwen3.6-27B-MTPLX-Optimized-Speed" --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.