← Back to community benchmarks

Qwen3.5-27BMTP

M5 Pro (20c) · 48 GB · 4bit · 2026-05-12
Performance
16k
tokens
456.6
PP tok/s
23.6
TG tok/s
35884
TTFT (ms)
20.1
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 48 GB
GPU Cores 20
Software
oMLX v0.3.9.dev1
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 433.3 28.2 16.1 GB view
4k 475.6 27.1 17.6 GB view
8k 475.0 25.7 18.6 GB view
16k 456.6 23.6 20.1 GB current
32k 416.7 21.4 23.1 GB view
64k 343.8 18.1 29.1 GB view
Batching Results
Batch Size TG tok/s Speedup
28.2 1.00×
31.6 1.12×
44.6 1.58×