← Back to community benchmarks

gemma-4-26B-A4B-it-uncensored-oQ5

M3 Ultra (80c) · 512 GB · 5bit · 2026-05-12
Performance
64k
tokens
1,653
PP tok/s
45.3
TG tok/s
39655
TTFT (ms)
20.8
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
32k 2,050 61.9 19.1 GB view
64k 1,653 45.3 20.8 GB current
128k 1,142 28.9 24.4 GB view