llama.cpp on the M5 Ultra 256GB
On the M5 Ultra 256GB, llama.cpp runs Qwen3.8 125B · 6B active at 8-bit at a median 17.0 tokens per second for generation, from 1 plain run (250K context).
Speed by model on the M5 Ultra 256GB
Plain runs only: one device, one request, no speculative decoding, the whole model in memory. Each row is one model size at one bit class, with the other engines' plain runs of the same model on this card beside it. Context length and build still differ between runs, and each side shows its own.
| Model | llama.cpp | Other engines |
|---|---|---|
| Qwen3.8 125B · 6B active 8-bit | 17.0 t/sone report, 250K context | no other engine reported |
3 llama.cpp reports on the M5 Ultra 256GB
Newest first. A run on several cards, with speculative decoding, with part of the model in system RAM or with batched requests says so, since each describes a different setup.
- Qwen3.8 125B · 6B active · IQ4_NL100.2 t/s(speculative decoding)
- GLM-5.3 320B · 18B active61.0 t/s(speculative decoding)
- Qwen3.8 125B · 6B active · Q8_017.0 t/s
Frequently asked
How fast is llama.cpp on the M5 Ultra 256GB?
On the M5 Ultra 256GB, llama.cpp runs Qwen3.8 125B · 6B active at 8-bit at a median 17.0 tokens per second for generation, from 1 plain run (250K context).