← Back to community benchmarks

gemma-4-26B-A4B-it-oQ8e

M2 Ultra (60c) · 192 GB · 8bit · 2026-04-17
Performance
128k
tokens
551.0
PP tok/s
9.2
TG tok/s
237895
TTFT (ms)
32.7
Peak mem (GB)
Hardware
Chip M2 Ultra (60c)
Memory 192 GB
GPU Cores 60
Software
oMLX v0.3.5
macOS macOS 26.3.1
Context 131,072
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 671.1 72.0 25.7 GB view
4k 864.1 66.5 26.2 GB view
8k 947.3 59.3 26.3 GB view
16k 929.8 45.6 26.7 GB view
32k 873.5 28.0 27.5 GB view
64k 755.1 16.6 29.2 GB view
128k 551.0 9.2 32.7 GB current
Batching Results
Batch Size TG tok/s Speedup
72.0 1.00×
102.6 1.42×
140.3 1.95×
184.9 2.57×