← Back to community benchmarks

gemma-4-26b-a4b-it

M3 Max (30c) · 36 GB · 4bit · 2026-04-03
Performance
8k
tokens
936.8
PP tok/s
42.7
TG tok/s
8745
TTFT (ms)
15.9
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.3.2
macOS macOS 26.3.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 942.9 50.4 15.3 GB view
4k 925.3 48.6 15.8 GB view
8k 936.8 42.7 15.9 GB current
16k 852.8 29.6 16.3 GB view
32k 728.1 17.3 17.1 GB view
64k 581.1 11.9 18.8 GB view
Batching Results
Batch Size TG tok/s Speedup
50.4 1.00×
60.8 1.21×