← Back to community benchmarks

gemma-4-26b-a4b-it

M3 Max (40c) · 128 GB · bf16 · 2026-05-20
Performance
32k
tokens
850.2
PP tok/s
27.8
TG tok/s
38542
TTFT (ms)
49.5
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 15.7.4
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 694.7 34.3 47.8 GB view
4k 911.3 33.6 48.5 GB view
8k 944.3 34.0 48.6 GB view
16k 925.8 31.8 49.1 GB view
32k 850.2 27.8 49.5 GB current
64k 648.9 19.8 51.2 GB view
128k 417.4 10.3 54.8 GB view
195k 315.3 9.4 58.5 GB view
Batching Results
Batch Size TG tok/s Speedup
34.3 1.00×
29.2 0.85×
52.9 1.54×
61.4 1.79×