← Back to community benchmarks

Qwen3.5-122B-A10B3.7bit-VL-v2

M2 Max (38c) · 96 GB · 3bit · 2026-04-05
Performance
1k
tokens
196.9
PP tok/s
33.8
TG tok/s
5201
TTFT (ms)
54.3
Peak mem (GB)
Hardware
Chip M2 Max (38c)
Memory 96 GB
GPU Cores 38
Software
oMLX v0.3.4
macOS macOS 15.7.4
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 196.9 33.8 54.3 GB current
4k 225.0 31.7 55.6 GB view
Batching Results
Batch Size TG tok/s Speedup
33.8 1.00×
46.4 1.37×
92.3 2.73×