← Back to community benchmarks

gemma-4-31B-it-uncensored-heretic-oQ4

M1 Max (24c) · 64 GB · 4bit · 2026-06-03
Performance
32k
tokens
81.0
PP tok/s
8.6
TG tok/s
404684
TTFT (ms)
23.4
Peak mem (GB)
Hardware
Chip M1 Max (24c)
Memory 64 GB
GPU Cores 24
Software
oMLX v0.4.0
macOS macOS 26.5.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 94.0 13.4 18.1 GB view
4k 93.3 12.7 19.9 GB view
8k 92.1 11.9 20.3 GB view
16k 88.9 10.6 21.0 GB view
32k 81.0 8.6 23.4 GB current
64k 70.2 6.0 28.1 GB view
128k 53.1 3.6 37.5 GB view
Batching Results
Batch Size TG tok/s Speedup
13.4 1.00×
15.0 1.12×
30.6 2.28×