← Back to community benchmarks

gemma-4-26b-a4b-it-mxfp8

M5 Max (40c) · 128 GB · 8bit · 2026-05-19
Performance
64k
tokens
2,352
PP tok/s
51.8
TG tok/s
27863
TTFT (ms)
29.3
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9rc1
macOS macOS 26.4.1
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,103 86.1 25.1 GB view
4k 2,754 82.9 25.7 GB view
8k 3,449 79.9 25.9 GB view
16k 3,413 74.1 26.3 GB view
32k 3,006 65.1 27.3 GB view
64k 2,352 51.8 29.3 GB current
128k 1,585 37.8 33.4 GB view
195k 1,204 28.7 38.3 GB view
Batching Results
Batch Size TG tok/s Speedup
86.1 1.00×
119.3 1.39×
168.7 1.96×
204.7 2.38×