← Back to community benchmarks

gemma-4-26B-A4B-it-TurboQuant

M4 (10c) · 32 GB · 4bit · 2026-05-26
Performance
32k
tokens
270.6
PP tok/s
20.4
TG tok/s
121090
TTFT (ms)
15.9
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.9
macOS macOS 26.3.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 361.2 34.2 14.2 GB view
4k 358.5 32.1 14.5 GB view
8k 329.5 28.2 14.8 GB view
16k 308.2 28.4 15.2 GB view
32k 270.6 20.4 15.9 GB current
64k 207.4 16.4 17.7 GB view