← Back to community benchmarks

Qwen3.5-27BMTP

M1 Max (32c) · 64 GB · 4bit · 2026-05-11
Performance
32k
tokens
83.5
PP tok/s
14.4
TG tok/s
392560
TTFT (ms)
22.8
Peak mem (GB)
Hardware
Chip M1 Max (32c)
Memory 64 GB
GPU Cores 32
Software
oMLX v0.3.9.dev1
macOS macOS 15.7.4
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 87.4 19.3 15.8 GB view
4k 89.0 18.5 17.3 GB view
8k 88.5 17.9 18.3 GB view
16k 87.0 16.3 19.8 GB view
32k 83.5 14.4 22.8 GB current
64k 76.1 12.3 28.8 GB view
Batching Results
Batch Size TG tok/s Speedup
19.3 1.00×
21.6 1.12×
23.7 1.23×