← Back to community benchmarks

Qwen3-4B-Instruct-2507

M1 Pro (14c) · 16 GB · 4bit · 2026-03-13
Performance
16k
tokens
172.2
PP tok/s
17.9
TG tok/s
95156
TTFT (ms)
5.0
Peak mem (GB)
Hardware
Chip M1 Pro (14c)
Memory 16 GB
GPU Cores 14
Software
oMLX v0.2.10
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
8k 207.2 27.5 3.9 GB view
16k 172.2 17.9 5.0 GB current