← Back to community benchmarks

Qwen3.5-35B-A3B-oQ3

M4 (10c) · 32 GB · 3bit · 2026-04-17
Performance
32k
tokens
346.7
PP tok/s
21.1
TG tok/s
94505
TTFT (ms)
18.9
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.6
macOS macOS 26.3.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 409.3 45.7 15.8 GB view
4k 457.4 43.1 16.6 GB view
8k 445.1 40.2 16.9 GB view
16k 409.1 33.3 17.6 GB view
32k 346.7 21.1 18.9 GB current
64k 258.3 12.8 21.5 GB view
Batching Results
Batch Size TG tok/s Speedup
45.7 1.00×
65.2 1.43×
75.6 1.65×
84.6 1.85×