// RUN — FITS

Can the Apple M2 Ultra 192GB run DeepSeek LLM 67B?

Yes — here's how fast and up to what context, with real measured numbers.

DeepSeek LLM 67B on the Apple M2 Ultra 192GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 114k 4.1-4.9 t/s ~0.1-0.1 (slow)
Q8 yes 318k 8.1-9.8 t/s ~0.1-0.1 (slow)
Q4 yes 408k 14.5-17.5 t/s ~0.2-0.2 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.

→ All models on the Apple M2 Ultra 192GB → DeepSeek LLM 67B specs → Other hardware

FAQ

Can the Apple M2 Ultra 192GB run DeepSeek LLM 67B?▶

Yes. At Q4 it generates ~14.5-17.5 tok/s and fits up to 408k context; beyond that it spills to SSD swap and slows to ~0.2-0.2 tok/s.

How much context fits?▶

At Q4, up to about 408k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.