← Back to community benchmarks

Qwen3.5-27B

M5 Pro (20c) · 64 GB · 4bit · 2026-03-16
Performance
16k
tokens
444.3
PP tok/s
14.6
TG tok/s
36879
TTFT (ms)
18.3
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 64 GB
GPU Cores 20
Software
oMLX v0.2.12
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
4k 478.9 17.8 17.0 GB view
8k 468.6 17.3 17.4 GB view
16k 444.3 14.6 18.3 GB current
32k 413.0 11.1 20.1 GB view
64k 344.6 8.0 23.6 GB view
128k 254.2 5.2 31.0 GB view
195k 197.6 3.8 38.6 GB view