← Back to community benchmarks

gemma-4-26b-a4b-it

M4 Max (32c) · 36 GB · 4bit · 2026-05-18
Performance
16k
tokens
1,087
PP tok/s
72.8
TG tok/s
15075
TTFT (ms)
15.6
Peak mem (GB)
Hardware
Chip M4 Max (32c)
Memory 36 GB
GPU Cores 32
Software
oMLX v0.3.8
macOS macOS 15.7.7
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 997.8 87.8 14.3 GB view
4k 1,108 84.2 14.9 GB view
8k 1,137 79.7 15.1 GB view
16k 1,087 72.8 15.6 GB current
32k 979.7 62.7 16.5 GB view
64k 806.7 48.3 18.5 GB view
128k 580.5 31.9 21.2 GB view
195k 439.4 23.6 25.0 GB view
Batching Results
Batch Size TG tok/s Speedup
87.8 1.00×
130.1 1.48×
167.0 1.90×