Benchmark result
NVIDIA-Nemotron-3-Super-120B-A12B-OptiQ-2bit on Apple M5 Max — 37.4 tok/s
Measured with oMLX on July 13, 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 3 runs of this model on Apple M5 Max · median 37.4 tok/s.
sort
- 141.6tok/s73.1oMLX
- 237.4tok/s72.1oMLXthis run
- 332.5tok/s47.9oMLX
Reproduce this run
tooloMLX
modelNVIDIA-Nemotron-3-Super-120B-A12B-OptiQ-2bit
context16,384
clientv0.4.22+97
Set those in oMLX, then:
llm-benchmark benchmark --model "NVIDIA-Nemotron-3-Super-120B-A12B-OptiQ-2bit" --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.