← Back to community benchmarks

Qwen3.5-4B

M1 (8c) · 16 GB · 4bit · 2026-03-13
Performance
4k
tokens
139.7
PP tok/s
20.2
TG tok/s
29324
TTFT (ms)
4.2
Peak mem (GB)
Hardware
Chip M1 (8c)
Memory 16 GB
GPU Cores 8
Software
oMLX v0.2.10
macOS macOS 26.3.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 144.0 22.1 3.9 GB view
4k 139.7 20.2 4.2 GB current
Batching Results
Batch Size TG tok/s Speedup
22.1 1.00×
33.4 1.51×
33.5 1.52×