← Back to community benchmarks

Gemma4-26B-MoE-Q8

M5 Max (40c) · 64 GB · 8bit · 2026-06-05
Performance
32k
tokens
2,514
PP tok/s
57.9
TG tok/s
13036
TTFT (ms)
28.0
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 64 GB
GPU Cores 40
Software
oMLX v0.4.1
macOS macOS 26.5.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,007 60.2 25.8 GB view
4k 2,770 81.8 26.5 GB view
8k 3,419 54.5 26.8 GB view
16k 3,112 73.5 27.1 GB view
32k 2,514 57.9 28.0 GB current
64k 2,003 51.2 30.0 GB view
Batching Results
Batch Size TG tok/s Speedup
60.2 1.00×
125.1 2.08×
173.7 2.89×
215.4 3.58×