← Back to community benchmarks

gemma-4-26b-a4b-it

M5 Max (40c) · 128 GB · 4bit · 2026-06-01
Performance
16k
tokens
3,640
PP tok/s
85.3
TG tok/s
4501
TTFT (ms)
22.2
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.12
macOS macOS 26.4
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,276 107.7 21.2 GB view
4k 3,410 95.8 21.6 GB view
8k 3,711 102.9 21.8 GB view
16k 3,640 85.3 22.2 GB current
32k 3,218 79.9 22.9 GB view
64k 2,436 57.7 24.6 GB view
128k 1,453 34.9 28.2 GB view
195k 897.2 24.9 31.9 GB view
Batching Results
Batch Size TG tok/s Speedup
107.7 1.00×
207.2 1.92×
351.5 3.26×
763.7 7.09×