← Back to community benchmarks

gemma-4-26b-a4b-it

M4 Max (40c) · 48 GB · 4bit · 2026-04-24
Performance
8k
tokens
1,305
PP tok/s
74.5
TG tok/s
6278
TTFT (ms)
14.8
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.3.7rc2
macOS macOS 26.4.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,039 96.7 14.3 GB view
4k 1,332 92.3 14.7 GB view
8k 1,305 74.5 14.8 GB current
16k 1,188 37.8 15.2 GB view
32k 1,035 14.8 16.0 GB view
64k 778.2 7.9 17.7 GB view
128k 485.6 3.8 21.2 GB view
195k 380.3 2.6 25.0 GB view