← Back to community benchmarks

gemma-4-26b-a4b-it

M4 Max (40c) · 128 GB · 4bit · 2026-07-23
Performance
64k
tokens
812.2
PP tok/s
54.1
TG tok/s
80691
TTFT (ms)
18.1
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.1
macOS macOS 26.5.2
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,574 99.9 14.3 GB view
4k 1,412 92.8 14.6 GB view
8k 1,247 90.6 15.0 GB view
16k 1,106 83.0 15.7 GB view
32k 979.9 71.0 16.8 GB view
64k 812.2 54.1 18.1 GB current
128k 505.5 34.1 21.4 GB view
195k 368.3 22.2 25.4 GB view
Batching Results
Batch Size TG tok/s Speedup
99.9 1.00×
156.8 1.57×
210.9 2.11×