← Back to community benchmarks

Qwen3.5-4B

M2 Ultra (76c) · 128 GB · 4bit · 2026-03-09
Performance
32k
tokens
1,247
PP tok/s
74.5
TG tok/s
26280
TTFT (ms)
6.9
Peak mem (GB)
Hardware
Chip M2 Ultra (76c)
Memory 128 GB
GPU Cores 76
Software
oMLX v0.2.6
macOS macOS 26.3
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,428 137.5 4.0 GB view
4k 1,440 128.2 4.2 GB view
8k 1,414 117.4 4.6 GB view
16k 1,355 98.7 5.3 GB view
32k 1,247 74.5 6.9 GB current
64k 1,051 50.4 8.6 GB view
Batching Results
Batch Size TG tok/s Speedup
137.5 1.00×
206.8 1.50×
255.6 1.86×
301.3 2.19×