← Back to community benchmarks

Qwen3.5-27B

M5 Max (32c) · 36 GB · 4bit · 2026-03-13
Performance
1k
tokens
703.4
PP tok/s
24.7
TG tok/s
1456
TTFT (ms)
16.6
Peak mem (GB)
Hardware
Chip M5 Max (32c)
Memory 36 GB
GPU Cores 32
Software
oMLX v0.2.10
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 703.4 24.7 16.6 GB current
4k 685.4 24.1 17.0 GB view
8k 657.2 22.9 17.4 GB view
16k 563.6 19.0 18.3 GB view
32k 521.3 13.7 20.1 GB view
Batching Results
Batch Size TG tok/s Speedup
24.7 1.00×
35.0 1.42×
51.0 2.06×