← Back to community benchmarks

gemma-4-31b-it

M3 Ultra (80c) · 512 GB · 6bit · 2026-04-18
Performance
16k
tokens
313.0
PP tok/s
16.3
TG tok/s
52352
TTFT (ms)
27.6
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 312.1 22.0 24.7 GB view
4k 323.0 20.8 26.5 GB view
8k 324.6 19.0 26.8 GB view
16k 313.0 16.3 27.6 GB current
32k 293.4 12.9 29.9 GB view
64k 252.5 8.8 34.6 GB view