← Back to community benchmarks

gemma-4-26b-a4b-it

M1 Ultra (48c) · 64 GB · 4bit · 2026-04-13
Performance
16k
tokens
685.8
PP tok/s
32.6
TG tok/s
23890
TTFT (ms)
15.2
Peak mem (GB)
Hardware
Chip M1 Ultra (48c)
Memory 64 GB
GPU Cores 48
Software
oMLX v0.3.4
macOS macOS 15.7.5
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 586.6 60.9 14.2 GB view
4k 679.7 55.1 14.7 GB view
8k 696.3 48.9 14.8 GB view
16k 685.8 32.6 15.2 GB current
32k 645.0 20.8 16.0 GB view
64k 562.6 13.1 17.7 GB view
Batching Results
Batch Size TG tok/s Speedup
60.9 1.00×
84.7 1.39×
113.6 1.87×
134.9 2.22×