← Back to community benchmarks

gemma-4-31b-abliterated

M3 Ultra (60c) · 256 GB · 4bit · 2026-04-06
Performance
8k
tokens
252.5
PP tok/s
24.8
TG tok/s
32440
TTFT (ms)
19.7
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.3.4
macOS macOS 26.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 251.5 31.1 17.6 GB view
4k 255.3 27.5 19.4 GB view
8k 252.5 24.8 19.7 GB current
16k 244.9 19.2 20.4 GB view
32k 227.7 11.1 22.8 GB view
64k 199.2 6.0 27.6 GB view
Batching Results
Batch Size TG tok/s Speedup
31.1 1.00×
47.2 1.52×
53.7 1.73×
60.4 1.94×