← Back to community benchmarks

gemma-4-26b-a4b-it

M5 Max (40c) · 128 GB · 4bit · 2026-05-22
Performance
195k
tokens
579.8
PP tok/s
20.5
TG tok/s
344963
TTFT (ms)
27.7
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9
macOS macOS 26.5
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,173 116.4 14.3 GB view
4k 3,323 110.0 14.9 GB view
8k 3,603 105.4 15.1 GB view
16k 3,220 95.6 15.6 GB view
32k 2,728 82.4 16.5 GB view
64k 2,179 61.9 18.5 GB view
128k 1,351 30.2 22.7 GB view
195k 579.8 20.5 27.7 GB current
Batching Results
Batch Size TG tok/s Speedup
116.4 1.00×
95.9 0.82×
123.5 1.06×
148.3 1.27×