← Back to community benchmarks

Qwen3.5-27B

M4 (10c) · 32 GB · 4bit · 2026-03-11
Performance
16k
tokens
57.5
PP tok/s
5.7
TG tok/s
285084
TTFT (ms)
18.2
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 63.4 6.8 16.6 GB view
4k 63.6 6.7 16.9 GB view
8k 62.1 6.5 17.3 GB view
16k 57.5 5.7 18.2 GB current
Batching Results
Batch Size TG tok/s Speedup
6.8 1.00×
13.5 1.99×