← Back to community benchmarks

Qwen3.6-27B-UD

M5 Max (40c) · 128 GB · 4bit · 2026-04-28
Performance
16k
tokens
719.4
PP tok/s
19.1
TG tok/s
22775
TTFT (ms)
28.5
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8rc1
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 600.0 21.5 25.2 GB view
4k 783.8 21.0 26.6 GB view
8k 752.7 20.6 27.2 GB view
16k 719.4 19.1 28.5 GB current
32k 654.2 17.5 31.0 GB view
64k 549.0 14.3 36.1 GB view
Batching Results
Batch Size TG tok/s Speedup
21.5 1.00×
23.6 1.10×
39.0 1.81×