← Back to community benchmarks

Qwen3.5-35B-A3B

M5 (10c) · 32 GB · 4bit · 2026-03-14
Performance
1k
tokens
1,041
PP tok/s
59.5
TG tok/s
984
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.2.11
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,041 59.5 20.0 GB current
4k 1,051 56.9 20.2 GB view
Batching Results
Batch Size TG tok/s Speedup
59.5 1.00×
93.0 1.56×
114.7 1.93×