← Back to community benchmarks

Qwen3-VL-8B-Instruct

M1 Pro (16c) · 16 GB · 4bit · 2026-03-09
Performance
8k
tokens
145.2
PP tok/s
22.6
TG tok/s
56417
TTFT (ms)
7.2
Peak mem (GB)
Hardware
Chip M1 Pro (16c)
Memory 16 GB
GPU Cores 16
Software
oMLX v0.2.6
macOS macOS 26.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 157.8 31.6 6.2 GB view
4k 153.8 27.1 6.6 GB view
8k 145.2 22.6 7.2 GB current
Batching Results
Batch Size TG tok/s Speedup
31.6 1.00×
35.3 1.12×