← Back to community benchmarks

gemma-4-31B-it

M5 Max (40c) · 128 GB · 8bit · 2026-05-21
Performance
16k
tokens
539.5
PP tok/s
13.6
TG tok/s
30369
TTFT (ms)
34.8
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9rc1
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 608.8 15.0 31.8 GB view
4k 651.5 14.8 33.8 GB view
8k 601.9 13.4 34.1 GB view
16k 539.5 13.6 34.8 GB current
32k 504.9 12.4 37.0 GB view
64k 426.3 10.5 41.8 GB view
Batching Results
Batch Size TG tok/s Speedup
15.0 1.00×
25.4 1.69×
34.7 2.31×