← Back to community benchmarks

Qwen3.5-35B-A3B

M3 Pro (18c) · 36 GB · 4bit · 2026-03-09
Performance
4k
tokens
685.1
PP tok/s
50.0
TG tok/s
5979
TTFT (ms)
20.2
Peak mem (GB)
Hardware
Chip M3 Pro (18c)
Memory 36 GB
GPU Cores 18
Software
oMLX v0.2.6
macOS macOS 26.3
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 656.7 54.4 20.1 GB view
4k 685.1 50.0 20.2 GB current
8k 663.8 47.3 20.4 GB view
16k 609.9 37.5 20.8 GB view
Batching Results
Batch Size TG tok/s Speedup
54.4 1.00×
84.9 1.56×
115.2 2.12×
144.9 2.66×