Benchmark result
K2-Horizon-MoVA-36B-A4B-MLX-4bit on Apple M5 Max — 47.8 tok/s
Measured with oMLX 0.31.3 on September 11, 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
3rd fastest of 3 runs of this model on Apple M5 Max · median 48.4 tok/s.
sort
- 149.1tok/s83.9oMLX · KV none
- 248.4tok/s84.7oMLX · KV none
- 347.8tok/s82.3oMLX · KV nonethis run
Reproduce this run
tooloMLX 0.31.3
modelK2-Horizon-MoVA-36B-A4B-MLX-4bit
context524,288
kv cachenone
thinkingon
efforthigh (model default)
think limitnone
samplingtemperature 1 · top_p 0.95 · top_k 20 · repetition_penalty 1
clientv0.4.59+97
Set those in oMLX, then:
llm-benchmark benchmark --model "K2-Horizon-MoVA-36B-A4B-MLX-4bit" --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.