// RUN ON APPLE

Run NeMO Megatron GPT 20B on Apple hardware

17 options ranked by estimated speed at Q4 / 8k context.

HardwareEst tok/sFits 8k?Max ctx~Price
Apple M2 Ultra 192GB โ†’ ~58.6 yes 611k ~$6,600
Apple M2 Ultra 64GB โ†’ ~58.6 yes 171k ~$4,000
Apple M1 Ultra 128GB โ†’ ~58.6 yes 391k ~$4,800
Apple M1 Ultra 64GB โ†’ ~58.6 yes 171k ~$3,700
Apple M4 Max 64GB โ†’ ~40 yes 171k ~$3,900
Apple M4 Max 36GB โ†’ ~40 yes 75k ~$3,000
Apple M4 Max 128GB โ†’ ~40 yes 391k ~$4,900
Apple M4 Max 48GB โ†’ ~40 yes 116k ~$3,300
Apple M1 Max 32GB โ†’ ~29.3 yes 61k ~$2,000
Apple M3 Max 48GB โ†’ ~29.3 yes 116k ~$3,200
Apple M3 Max 64GB โ†’ ~29.3 yes 171k ~$3,700
Apple M3 Max 128GB โ†’ ~29.3 yes 391k ~$4,700
Apple M1 Max 64GB โ†’ ~29.3 yes 171k ~$2,600
Apple M2 Max 32GB โ†’ ~29.3 yes 61k ~$2,200
Apple M2 Max 64GB โ†’ ~29.3 yes 171k ~$2,800
Apple M2 Max 96GB โ†’ ~29.3 yes 281k ~$3,300
Apple M3 Max 36GB โ†’ ~22 yes 75k ~$2,900

Estimate at Q4 (memory-bound). Click any option for its full per-hardware estimate and measured numbers.

FAQ

What Apple hardware runs NeMO Megatron GPT 20B?โ–ถ

On GenAIList, 17 Apple options can run NeMO Megatron GPT 20B at Q4 โ€” the fastest is the Apple M2 Ultra 192GB at ~58.6 tok/s.

How fast is NeMO Megatron GPT 20B on Apple hardware?โ–ถ

Estimated generation speed ranges from ~22 to ~58.6 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.