← Back to community benchmarks

gemma-4-31B-it-heretic-oQ5

M4 Max (40c) · 48 GB · 5bit · 2026-04-18
Performance
16k
tokens
173.0
PP tok/s
17.2
TG tok/s
94702
TTFT (ms)
26.7
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.3.6
macOS macOS 26.5
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 195.5 19.6 21.7 GB view
4k 183.0 19.1 24.2 GB view
8k 170.3 18.6 25.1 GB view
16k 173.0 17.2 26.7 GB current
Batching Results
Batch Size TG tok/s Speedup
19.6 1.00×
27.2 1.39×
29.7 1.52×
31.3 1.60×