← Back to community benchmarks

Llama-3.1-8B-Instruct

M2 (10c) · 16 GB · 4bit · 2026-04-24
Performance
4k
tokens
184.5
PP tok/s
17.9
TG tok/s
22201
TTFT (ms)
5.3
Peak mem (GB)
Hardware
Chip M2 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 188.3 20.0 4.9 GB view
4k 184.5 17.9 5.3 GB current
Batching Results
Batch Size TG tok/s Speedup
20.0 1.00×
35.0 1.75×
70.2 3.51×