← Back to community benchmarks

gemma-4-26B-A4B-it-oQ3

M2 (10c) · 24 GB · 3bit · 2026-05-28
Performance
64k
tokens
133.3
PP tok/s
10.3
TG tok/s
491594
TTFT (ms)
14.8
Peak mem (GB)
Hardware
Chip M2 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.12
macOS macOS 26.5
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 291.5 27.1 11.4 GB view
4k 299.6 26.3 11.7 GB view
8k 277.0 23.8 11.9 GB view
16k 192.0 15.3 12.3 GB view
32k 152.8 12.6 13.1 GB view
64k 133.3 10.3 14.8 GB current
Batching Results
Batch Size TG tok/s Speedup
27.1 1.00×
24.9 0.92×
28.6 1.06×
33.0 1.22×