← Back to community benchmarks

Llama-3.3-70B-Instruct

M2 Max (30c) · 64 GB · 4bit · 2026-03-10
Performance
1k
tokens
58.3
PP tok/s
8.5
TG tok/s
17576
TTFT (ms)
37.8
Peak mem (GB)
Hardware
Chip M2 Max (30c)
Memory 64 GB
GPU Cores 30
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 58.3 8.5 37.8 GB current
4k 54.3 7.9 39.0 GB view
Batching Results
Batch Size TG tok/s Speedup
8.5 1.00×
11.9 1.40×
12.3 1.45×