← Back to community benchmarks

Qwen3.5-27B

M1 Max (32c) · 64 GB · 4bit · 2026-06-04
User commentregular 4bit
Performance
1k
tokens
80.4
PP tok/s
13.5
TG tok/s
12735
TTFT (ms)
15.8
Peak mem (GB)
Hardware
Chip M1 Max (32c)
Memory 64 GB
GPU Cores 32
Software
oMLX v0.4.1
macOS macOS 26.4
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 80.4 13.5 15.8 GB current
4k 66.1 11.9 17.3 GB view
Batching Results
Batch Size TG tok/s Speedup
13.5 1.00×
15.4 1.14×
18.2 1.35×