Benchmark result
Ornith-1.5-35B-A3B-oQ4e-mtp on Apple M4 Pro — 73.9 tok/s
Measured with oMLX manual on October 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
2nd fastest of 11 runs of this model on Apple M4 Pro · median 71.1 tok/s.
sort
- 173.9tok/s82.3oMLX · oQ4e · KV none
- 273.9tok/s86.8oMLXthis run
- 373.1tok/s80.2oMLX · oQ4e · KV none
- 473.0tok/s84.0oMLX
- 571.9tok/s83.3oMLX · KV none
- 671.1tok/s79.8oMLX · oQ4e · KV none
- 770.7tok/s82.7oMLX · oQ4e · KV none
- 870.6tok/s86.1oMLX · KV none
- 970.1tok/s85.0oMLX · KV none
- 1069.4tok/s86.2oMLX · KV none
- 1168.7tok/s81.1oMLX · KV none
Ornith-1.5-35B-A3B-oQ4e-mtp on other hardware
Reproduce this run
tooloMLX manual
modelOrnith-1.5-35B-A3B-oQ4e-mtp
context262,144
thinkingon
clientv0.4.93+102
Set those in oMLX, then:
llm-benchmark benchmark --model "Ornith-1.5-35B-A3B-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.