← Back to community benchmarks

llama-3.1-8b-instruct

M4 (10c) · 16 GB · 4bit · 2026-03-10
Performance
1k
tokens
236.2
PP tok/s
20.5
TG tok/s
4339
TTFT (ms)
4.9
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 15.6.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 236.2 20.5 4.9 GB current
4k 27.2 1.6 5.3 GB view
Batching Results
Batch Size TG tok/s Speedup
20.5 1.00×
38.3 1.87×
54.2 2.64×