← Back to community benchmarks

Llama-3.2-3B-Instruct

M3 Max (30c) · 36 GB · unknown · 2026-03-21
Performance
1k
tokens
1,258
PP tok/s
40.2
TG tok/s
815
TTFT (ms)
6.5
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.2.12
macOS macOS 26.3
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,258 40.2 6.5 GB current
4k 1,342 38.3 6.8 GB view
Batching Results
Batch Size TG tok/s Speedup
40.2 1.00×
45.0 1.12×
84.3 2.10×