← Back to community benchmarks

gemma-4-31b-it

M3 Ultra (60c) · 256 GB · 8bit · 2026-05-27
Performance
128k
tokens
150.0
PP tok/s
8.1
TG tok/s
873649
TTFT (ms)
57.9
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.3.11
macOS macOS 26.5
Context 131,072
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 246.4 18.3 31.8 GB view
4k 252.8 17.8 34.5 GB view
8k 249.8 17.3 35.0 GB view
16k 240.1 16.3 36.9 GB view
32k 222.3 14.6 39.0 GB view
64k 193.2 12.1 45.8 GB view
128k 150.0 8.1 57.9 GB current
Batching Results
Batch Size TG tok/s Speedup
18.3 1.00×
30.1 1.64×
32.0 1.75×