← Back to community benchmarks

Qwen3.5-27B

M4 (10c) · 32 GB · 4bit · 2026-04-18
Performance
4k
tokens
40.4
PP tok/s
5.3
TG tok/s
101373
TTFT (ms)
17.2
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.5.dev2
macOS macOS 15.7.4
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 40.3 5.4 15.8 GB view
4k 40.4 5.3 17.2 GB current
Batching Results
Batch Size TG tok/s Speedup
5.4 1.00×
8.5 1.57×
16.9 3.13×