← Back to community benchmarks

Qwen3.6-27B

M3 Max (40c) · 48 GB · 4bit · 2026-04-29
Performance
1k
tokens
197.1
PP tok/s
22.0
TG tok/s
5195
TTFT (ms)
15.8
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.3.8rc1
macOS macOS 15.7.5
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 197.1 22.0 15.8 GB current
8k 171.7 20.4 17.9 GB view
Batching Results
Batch Size TG tok/s Speedup
22.0 1.00×
36.6 1.66×
43.5 1.98×
45.3 2.06×