← Back to community benchmarks

Qwen3.5-4B

M1 (7c) · 16 GB · 4bit · 2026-03-10
Performance
1k
tokens
123.0
PP tok/s
19.9
TG tok/s
8324
TTFT (ms)
3.9
Peak mem (GB)
Hardware
Chip M1 (7c)
Memory 16 GB
GPU Cores 7
Software
oMLX v0.2.6
macOS macOS 15.2
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 123.0 19.9 3.9 GB current
4k 120.7 20.0 4.2 GB view
Batching Results
Batch Size TG tok/s Speedup
19.9 1.00×
28.3 1.42×
29.1 1.46×