← Back to community benchmarks

Qwen3.5-122B-A10B

M3 Max (40c) · 128 GB · 4bit · 2026-03-16
Performance
32k
tokens
313.5
PP tok/s
18.5
TG tok/s
104526
TTFT (ms)
69.1
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.2.13
macOS macOS 15.7.3
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 331.4 34.7 66.3 GB view
4k 362.6 34.1 66.7 GB view
8k 364.4 31.4 67.0 GB view
32k 313.5 18.5 69.1 GB current
Batching Results
Batch Size TG tok/s Speedup
34.7 1.00×
43.1 1.24×
38.9 1.12×
40.9 1.18×