← Back to community benchmarks

gemma-4-26b-a4b

M5 Max (40c) · 128 GB · 8bit · 2026-04-06
Performance
195k
tokens
1,195
PP tok/s
11.0
TG tok/s
167383
TTFT (ms)
36.4
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.4
macOS macOS 26.4
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,875 83.9 25.7 GB view
4k 2,842 77.9 26.2 GB view
8k 3,160 72.6 26.3 GB view
16k 3,129 62.8 26.7 GB view
32k 2,763 50.2 27.5 GB view
64k 2,159 28.2 29.2 GB view
128k 1,507 15.8 32.7 GB view
195k 1,195 11.0 36.4 GB current
Batching Results
Batch Size TG tok/s Speedup
83.9 1.00×
123.3 1.47×
169.4 2.02×
214.4 2.56×