← Back to community benchmarks

Qwen2.5-14B-Instruct

M3 Pro (18c) · 18 GB · 4bit · 2026-03-11
Performance
1k
tokens
183.4
PP tok/s
14.8
TG tok/s
5583
TTFT (ms)
8.5
Peak mem (GB)
Hardware
Chip M3 Pro (18c)
Memory 18 GB
GPU Cores 18
Software
oMLX v0.2.8
macOS macOS 26.4
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 183.4 14.8 8.5 GB current
4k 175.8 13.4 9.2 GB view
8k 157.7 11.8 10.0 GB view
Batching Results
Batch Size TG tok/s Speedup
14.8 1.00×
28.4 1.92×
42.0 2.84×
41.4 2.80×