← Back to community benchmarks

gemma-4-26b-a4b-it

M5 Max (40c) · 128 GB · 4bit · 2026-04-02
Performance
16k
tokens
2,715
PP tok/s
57.1
TG tok/s
6035
TTFT (ms)
16.0
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.1
macOS macOS 26.4
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 3,359 78.4 15.1 GB view
4k 2,249 74.8 15.4 GB view
8k 2,670 66.0 15.6 GB view
16k 2,715 57.1 16.0 GB current
32k 2,565 40.9 16.8 GB view
64k 2,058 23.8 18.6 GB view
Batching Results
Batch Size TG tok/s Speedup
78.4 1.00×
96.7 1.23×
139.1 1.77×
184.8 2.36×