← Back to community benchmarks

Qwen3.5-35B-A3B

M1 Max (24c) · 32 GB · 4bit · 2026-03-10
Performance
16k
tokens
355.3
PP tok/s
36.8
TG tok/s
46108
TTFT (ms)
20.8
Peak mem (GB)
Hardware
Chip M1 Max (24c)
Memory 32 GB
GPU Cores 24
Software
oMLX v0.2.6
macOS macOS 26.3
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 422.8 55.7 20.0 GB view
4k 430.4 52.2 20.2 GB view
8k 394.3 47.1 20.4 GB view
16k 355.3 36.8 20.8 GB current
Batching Results
Batch Size TG tok/s Speedup
55.7 1.00×
77.6 1.39×
84.3 1.51×