Benchmark result
Qwen3.8-27B-oQ8-mtp on Apple M5 Max — 36.0 tok/s
Measured with oMLX 0.6.0.dev1 on August 15, 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.
How this run compares
2nd fastest of 4 runs of this model on Apple M5 Max · median 35.1 tok/s.
sort
- 139.2tok/s80.2oMLX
- 236.0tok/s75.9oMLXthis run
- 334.2tok/s73.1oMLX
- 433.1tok/s81.6oMLX
Reproduce this run
tooloMLX 0.6.0.dev1
modelQwen3.8-27B-oQ8-mtp
context65,536
clientv0.4.36+97
Set those in oMLX, then:
llm-benchmark benchmark --model "Qwen3.8-27B-oQ8-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.