← Back to community benchmarks

Qwen3.6-27B

M3 Max (30c) · 36 GB · 4bit · 2026-04-26
User commentmlx-community/Qwen3.6-27B-4bit
Performance
1k
tokens
159.4
PP tok/s
17.3
TG tok/s
6425
TTFT (ms)
15.8
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.3.7rc1
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 159.4 17.3 15.8 GB current
4k 151.7 16.7 17.3 GB view
Batching Results
Batch Size TG tok/s Speedup
17.3 1.00×
32.5 1.88×
34.1 1.97×