← Back to community benchmarks

Qwen3.6-35B-A3B-MTP

M5 Max (40c) · 128 GB · 8bit · 2026-07-24
User commentmtp
Performance
32k
tokens
1,889
PP tok/s
85.6
TG tok/s
17344
TTFT (ms)
36.8
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.3
macOS macOS 26.5.2
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,348 100.7 35.5 GB view
4k 3,745 98.3 36.4 GB view
8k 4,088 96.0 36.7 GB view
16k 2,447 92.0 36.5 GB view
32k 1,889 85.6 36.8 GB current
64k 1,177 74.6 37.4 GB view
128k 692.8 59.6 38.7 GB view
195k 494.6 50.7 40.1 GB view
Batching Results
Batch Size TG tok/s Speedup
100.7 1.00×
210.0 2.09×
419.9 4.17×
840.8 8.35×