← Back to community benchmarks

Qwen3.5-122B-A10Bvision-4.7-bit

M3 Max (40c) · 128 GB · 4bit · 2026-04-05
Performance
4k
tokens
502.5
PP tok/s
39.5
TG tok/s
8152
TTFT (ms)
68.5
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.4
macOS macOS 26.4
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 426.7 41.3 67.2 GB view
4k 502.5 39.5 68.5 GB current
Batching Results
Batch Size TG tok/s Speedup
41.3 1.00×
61.8 1.50×
73.9 1.79×