MLX on the M5 Ultra 256GB
llamaperf has 6 MLX reports on the M5 Ultra 256GB, none of them a plain single-card run with a model size and quant, so there is no median to quote. The reports are listed below with how each one ran.
6 MLX 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.
- Gemma 4 · BF1613.9 t/s
- Qwen3.6 · Q4130.4 t/s
- Qwen3.5 · Q474.7 t/s
- Qwen3 · Q467.0 t/s
- Qwen3.8 27B · 4bit47.6 t/s
- Qwen3.8 125B · 6B active · oQ4e108.0 t/s(speculative decoding)
Frequently asked
How fast is MLX on the M5 Ultra 256GB?
llamaperf has 6 MLX reports on the M5 Ultra 256GB, none of them a plain single-card run with a model size and quant, so there is no median to quote. The reports are listed below with how each one ran.