← Back to community benchmarks

gemma-4-26B-A4B-it-oQ5

M5 Max (40c) · 128 GB · 5bit · 2026-05-04
Performance
8k
tokens
3,502
PP tok/s
92.3
TG tok/s
2339
TTFT (ms)
18.6
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 26.5
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,688 100.9 17.8 GB view
4k 2,929 94.8 18.5 GB view
8k 3,502 92.3 18.6 GB current
16k 3,478 84.0 19.1 GB view
32k 3,123 72.8 20.1 GB view
64k 2,506 56.8 22.1 GB view
Batching Results
Batch Size TG tok/s Speedup
100.9 1.00×
148.9 1.48×
206.0 2.04×