← Back to community benchmarks

gemma-4-26b-a4b-it

M2 Max (38c) · 32 GB · 4bit · 2026-05-07
Performance
64k
tokens
505.4
PP tok/s
34.7
TG tok/s
129672
TTFT (ms)
18.5
Peak mem (GB)
Hardware
Chip M2 Max (38c)
Memory 32 GB
GPU Cores 38
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 553.7 70.8 14.3 GB view
4k 616.2 66.9 14.9 GB view
8k 629.1 62.8 15.1 GB view
16k 614.4 56.2 15.6 GB view
32k 574.7 46.1 16.5 GB view
64k 505.4 34.7 18.5 GB current
128k 397.8 22.3 21.2 GB view
Batching Results
Batch Size TG tok/s Speedup
70.8 1.00×
96.1 1.36×
192.1 2.71×