← Back to community benchmarks

gemma-4-26B-A4B-it-oQ4

M4 (10c) · 32 GB · 4bit · 2026-05-06
Performance
1k
tokens
366.2
PP tok/s
31.6
TG tok/s
2797
TTFT (ms)
14.8
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.8
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 366.2 31.6 14.8 GB current
4k 375.4 31.0 15.5 GB view
8k 368.6 28.5 15.6 GB view
16k 804.6 11.6 16.5 GB view
32k 726.9 11.3 17.4 GB view
64k 591.3 10.5 18.3 GB view
Batching Results
Batch Size TG tok/s Speedup
31.6 1.00×
46.5 1.47×
55.3 1.75×
63.6 2.01×