← Back to community benchmarks

gemma-4-26B-A4B-it

M3 Max (30c) · 36 GB · 6bit · 2026-05-07
Performance
32k
tokens
765.0
PP tok/s
43.8
TG tok/s
42835
TTFT (ms)
22.2
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.3.8
macOS macOS 15.7.5
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 791.6 56.1 20.0 GB view
4k 895.6 54.6 20.7 GB view
8k 898.1 52.0 20.8 GB view
16k 845.5 48.7 21.3 GB view
32k 765.0 43.8 22.2 GB current
64k 643.8 35.5 24.3 GB view
128k 435.7 25.3 27.0 GB view
Batching Results
Batch Size TG tok/s Speedup
56.1 1.00×
76.7 1.37×
93.9 1.67×