← Back to community benchmarks

Qwen3.5-27B

M1 Ultra (48c) · 64 GB · 8bit · 2026-04-16

Performance

16k

tokens

128.2

PP tok/s

15.4

TG tok/s

127828

TTFT (ms)

31.7

Peak mem (GB)

Hardware

Chip M1 Ultra (48c)

Memory 64 GB

GPU Cores 48

Software

oMLX v0.3.5

macOS macOS 26.5

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
4k	84.4	17.5	29.8 GB	view
8k	130.2	16.7	30.4 GB	view
16k	128.2	15.4	31.7 GB	current
32k	123.2	13.0	34.2 GB	view