← Back to community benchmarks

Qwen3.6-27BoQ4

M3 Max (40c) · 128 GB · 4bit · 2026-04-23
Performance
32k
tokens
172.8
PP tok/s
12.1
TG tok/s
189580
TTFT (ms)
41.3
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.7
macOS macOS 26.4.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 214.9 22.1 35.5 GB view
4k 210.1 21.2 36.9 GB view
8k 193.3 18.8 37.5 GB view
16k 198.1 17.1 38.8 GB view
32k 172.8 12.1 41.3 GB current
Batching Results
Batch Size TG tok/s Speedup
22.1 1.00×
30.6 1.38×
38.2 1.73×