← Back to community benchmarks

gemma-4-31b-it-mxfp8

M3 Ultra (60c) · 256 GB · 8bit · 2026-05-21
Performance
32k
tokens
220.4
PP tok/s
15.1
TG tok/s
148674
TTFT (ms)
38.3
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.3.9
macOS macOS 26.5
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 245.1 18.4 30.9 GB view
4k 248.7 18.0 33.6 GB view
16k 237.5 16.6 35.3 GB view
32k 220.4 15.1 38.3 GB current
64k 192.3 12.6 44.2 GB view
128k 149.8 9.4 57.0 GB view
Batching Results
Batch Size TG tok/s Speedup
18.4 1.00×
23.9 1.30×
26.5 1.44×