← Back to community benchmarks

Qwen3.5-35B-A3B

M4 Pro (20c) · 48 GB · 6bit · 2026-03-10
Performance
8k
tokens
839.2
PP tok/s
55.7
TG tok/s
9762
TTFT (ms)
28.5
Peak mem (GB)
Hardware
Chip M4 Pro (20c)
Memory 48 GB
GPU Cores 20
Software
oMLX v0.2.6
macOS macOS 26.3
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 850.8 61.7 28.1 GB view
4k 864.0 55.4 28.3 GB view
8k 839.2 55.7 28.5 GB current
16k 775.0 49.4 28.9 GB view
32k 669.2 39.5 29.7 GB view
Batching Results
Batch Size TG tok/s Speedup
61.7 1.00×
95.7 1.55×
112.7 1.83×
131.5 2.13×