← Back to community benchmarks

Qwen3.6-35B-A3B-TurboQuant

M5 Pro (20c) · 64 GB · 4bit · 2026-05-10
Performance
1k
tokens
1,404
PP tok/s
103.5
TG tok/s
730
TTFT (ms)
20.1
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 64 GB
GPU Cores 20
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,404 103.5 20.1 GB current
4k 2,242 99.6 20.9 GB view
8k 2,315 94.7 21.2 GB view
16k 4,166 96.0 22.5 GB view
32k 4,362 92.1 23.0 GB view
64k 3,987 82.8 24.0 GB view
Batching Results
Batch Size TG tok/s Speedup
103.5 1.00×
155.2 1.50×
309.4 2.99×
620.7 6.00×