← Back to community benchmarks

Qwen3.6-27B-OptiQ

M5 Max (40c) · 128 GB · 4bit · 2026-06-20
Performance
195k
tokens
320.3
PP tok/s
16.1
TG tok/s
624346
TTFT (ms)
51.2
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.4.4
macOS macOS 26.5.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 750.8 26.9 19.3 GB view
4k 800.6 25.9 20.7 GB view
8k 741.4 26.1 21.4 GB view
16k 669.7 25.2 22.6 GB view
32k 632.1 24.4 25.1 GB view
64k 544.2 22.0 30.1 GB view
128k 398.6 18.1 40.5 GB view
195k 320.3 16.1 51.2 GB current
Batching Results
Batch Size TG tok/s Speedup
26.9 1.00×
49.9 1.86×
64.6 2.40×