← Back to community benchmarks

gemma-4-26b-a4b-it-UD

M5 Max (40c) · 128 GB · 4bit · 2026-05-12
Performance
16k
tokens
3,675
PP tok/s
89.8
TG tok/s
4458
TTFT (ms)
16.6
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9.dev1
macOS macOS 26.5
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,235 105.5 15.2 GB view
4k 3,348 101.0 15.7 GB view
8k 3,721 96.8 16.2 GB view
16k 3,675 89.8 16.6 GB current
32k 3,250 77.2 17.4 GB view
64k 2,574 61.5 19.4 GB view
Batching Results
Batch Size TG tok/s Speedup
105.5 1.00×
152.5 1.45×
217.0 2.06×