← Back to community benchmarks

Qwen3.5-122B-A10B

M5 Max (40c) · 128 GB · 6bit · 2026-04-02
Performance
16k
tokens
1,073
PP tok/s
38.8
TG tok/s
15265
TTFT (ms)
97.3
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.0
macOS macOS 26.4
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 955.7 51.1 94.0 GB view
4k 1,203 48.5 95.5 GB view
8k 1,175 44.6 96.1 GB view
16k 1,073 38.8 97.3 GB current
Batching Results
Batch Size TG tok/s Speedup
51.1 1.00×
61.7 1.21×