← Back to community benchmarks

gemma-4-31B-it-OptiQ

M5 Max (40c) · 128 GB · 4bit · 2026-05-22
Performance
16k
tokens
485.5
PP tok/s
16.4
TG tok/s
33745
TTFT (ms)
26.6
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9rc1
macOS macOS 26.5
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 644.7 21.3 22.3 GB view
4k 587.1 19.7 24.9 GB view
8k 538.8 17.9 25.5 GB view
16k 485.5 16.4 26.6 GB current
32k 395.8 8.1 29.6 GB view
64k 204.3 7.0 35.6 GB view
128k 169.1 5.4 47.6 GB view
195k 130.0 4.2 62.6 GB view
Batching Results
Batch Size TG tok/s Speedup
21.3 1.00×
16.5 0.77×
22.0 1.03×
25.4 1.19×