← Back to community benchmarks

gemma-4-26b-a4b

M2 Ultra (76c) · 192 GB · 4bit · 2026-05-14
Performance
16k
tokens
1,102
PP tok/s
63.6
TG tok/s
14867
TTFT (ms)
15.6
Peak mem (GB)
Hardware
Chip M2 Ultra (76c)
Memory 192 GB
GPU Cores 76
Software
oMLX v0.3.8
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 851.0 79.3 14.3 GB view
4k 1,072 76.1 14.9 GB view
8k 1,114 71.3 15.1 GB view
16k 1,102 63.6 15.6 GB current
32k 1,037 51.1 16.5 GB view
64k 877.4 33.6 18.5 GB view
Batching Results
Batch Size TG tok/s Speedup
79.3 1.00×
118.9 1.50×
235.4 2.97×
409.9 5.17×