← Back to community benchmarks

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

M5 (10c) · 32 GB · 3bit · 2026-04-26
Performance
16k
tokens
1,002
PP tok/s
30.2
TG tok/s
16353
TTFT (ms)
12.3
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.7
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 907.1 48.6 11.4 GB view
4k 1,025 45.8 11.8 GB view
8k 1,059 38.5 11.9 GB view
16k 1,002 30.2 12.3 GB current
32k 864.6 20.5 13.1 GB view
Batching Results
Batch Size TG tok/s Speedup
48.6 1.00×
66.0 1.36×
77.8 1.60×