← Back to community benchmarks

Qwen3.5-35B-A3B-oQ3e

M4 (10c) · 32 GB · 3bit · 2026-04-02
Performance
16k
tokens
402.7
PP tok/s
33.6
TG tok/s
40690
TTFT (ms)
17.6
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.0
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 398.4 45.4 15.8 GB view
4k 446.6 42.9 16.6 GB view
8k 437.1 40.3 17.0 GB view
16k 402.7 33.6 17.6 GB current
32k 343.1 21.1 19.0 GB view
64k 257.3 13.4 21.7 GB view
Batching Results
Batch Size TG tok/s Speedup
45.4 1.00×
64.9 1.43×
69.9 1.54×
77.1 1.70×