← Back to community benchmarks

gemma-4-31B-it

M5 Max (40c) · 128 GB · 6bit · 2026-04-30
Performance
16k
tokens
515.4
PP tok/s
14.5
TG tok/s
31789
TTFT (ms)
51.7
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.7
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 618.7 16.0 48.7 GB view
4k 679.6 15.7 50.8 GB view
8k 655.5 15.3 51.1 GB view
16k 515.4 14.5 51.7 GB current
32k 479.9 13.2 53.9 GB view
64k 406.0 11.2 58.7 GB view
128k 295.1 8.5 68.2 GB view
195k 227.5 6.8 78.2 GB view
Batching Results
Batch Size TG tok/s Speedup
16.0 1.00×
27.5 1.72×
36.1 2.26×
70.9 4.43×